[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15873022#comment-15873022 ] Kevin Liew commented on PHOENIX-476: Hi [~tdsilva], support for sequence expressions is planned for PHOENIX-3425. We do expect this error when parsing sequence values as default expression: {code:java} CANNOT_CREATE_DEFAULT(1031, "42Y90", "Cannot create column with a stateful default value.") {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: New Feature >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15867069#comment-15867069 ] Thomas D'Silva commented on PHOENIX-476: [~kliew] Can you use a sequence (NEXT VALUE FOR) as a default expression ? I get a parser error when I tried to use a sequence. {code} "CREATE TABLE IF NOT EXISTS " + sharedTable1 + " (" + "pk1 INTEGER NOT NULL, " + "pk2 INTEGER NOT NULL DEFAULT NEXT VALUE FOR my_seq, " + "CONSTRAINT NAME_PK PRIMARY KEY (pk1, pk2))" {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: New Feature >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.15#6346)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15618521#comment-15618521 ] Kevin Liew commented on PHOENIX-476: Thanks for the walkthrough James. It was a good overview of the Phoenix codebase and feature-set. I created PHOENIX-3425 as a follow-up for stateful default values. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15610494#comment-15610494 ] Hudson commented on PHOENIX-476: FAILURE: Integrated in Jenkins build Phoenix-master #1451 (See [https://builds.apache.org/job/Phoenix-master/1451/]) PHOENIX-476 Support declaration of DEFAULT in CREATE statement (Kevin (jamestaylor: rev 5744c6f0eb7119e20122a24e93c0a0c6e98e9662) * (add) phoenix-core/src/main/java/org/apache/phoenix/schema/DelegateSQLException.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/compile/UpsertCompiler.java * (edit) phoenix-core/src/test/java/org/apache/phoenix/compile/WhereCompilerTest.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/exception/SQLExceptionCode.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/compile/CreateTableCompiler.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/coprocessor/MetaDataEndpointImpl.java * (edit) phoenix-core/src/test/java/org/apache/phoenix/compile/QueryCompilerTest.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/coprocessor/UngroupedAggregateRegionObserver.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/parse/CreateTableStatement.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/schema/types/PBinary.java * (add) phoenix-core/src/main/java/org/apache/phoenix/expression/function/DefaultValueExpression.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/parse/UseSchemaStatement.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/execute/TupleProjector.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/parse/CreateSchemaStatement.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/parse/ParseNodeFactory.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/schema/ColumnRef.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/parse/ColumnDef.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/schema/MetaDataClient.java * (edit) phoenix-core/src/main/antlr3/PhoenixSQL.g * (add) phoenix-core/src/it/java/org/apache/phoenix/end2end/DefaultColumnValueIT.java * (edit) phoenix-core/src/it/java/org/apache/phoenix/iterate/MockResultIterator.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/schema/PTableImpl.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/expression/ExpressionType.java * (edit) phoenix-core/src/main/java/org/apache/phoenix/schema/PMetaDataImpl.java > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15610340#comment-15610340 ] James Taylor commented on PHOENIX-476: -- Pushed to 4.x and master branch. Great work, [~kliew]! > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15610032#comment-15610032 ] Hadoop QA commented on PHOENIX-476: --- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12835425/PHOENIX-476.12.patch against master branch at commit 202b8eb1eda29db01006cdeefd8199f0bd360692. ATTACHMENT ID: 12835425 {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 6 new or modified tests. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:red}-1 javadoc{color}. The javadoc tool appears to have generated 43 warning messages. {color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings. {color:red}-1 lineLengths{color}. The patch introduces the following lines longer than 100: +conn.createStatement().executeQuery("SELECT c1 FROM " + sharedTable2 + " WHERE c1 = 10"); +rs = conn.createStatement().executeQuery("SELECT c4 FROM " + sharedTable2 + " WHERE c4 = 'ABCD'"); +: c=column_name dt=identifier (LPAREN l=NUMBER (COMMA s=NUMBER)? RPAREN)? ar=ARRAY? (lsq=LSQUARE (a=NUMBER)? RSQUARE)? (nn=NOT? n=NULL)? (DEFAULT df=expression)? (pk=PRIMARY KEY (order=ASC|order=DESC)? rr=ROW_TIMESTAMP?)? +if (columnDef.isRowTimestamp() || ( pkConstraint != null && pkConstraint.isColumnRowTimestamp(columnDef.getColumnDefName( { +LiteralExpression defaultValue = ExpressionUtil.getConstantExpression(defaultExpression, ptr); +targetType.coerceBytes(context.getTempPtr(), defaultValue.getValue(), sourceType, +throw new DelegateSQLException(sqlE, ". DEFAULT " + SQLExceptionInfo.COLUMN_NAME + "=" + columnDef.getColumnDefName().getColumnName()); + SQLExceptionCode.DATA_EXCEEDS_MAX_CAPACITY).setColumnName(columnDef.getColumnDefName().getColumnName()) +return client.createTable(finalCreate, splits, parent, viewStatement, viewType, viewColumnConstants, isViewColumnReferenced); +protected RegionScanner doPostScannerOpen(final ObserverContext c, final Scan scan, final RegionScanner s) throws IOException, SQLException { {color:red}-1 core tests{color}. The patch failed these unit tests: ./phoenix-core/target/failsafe-reports/TEST-org.apache.phoenix.end2end.AlterTableIT {color:red}-1 core zombie tests{color}. There are 3 zombie test(s): at org.apache.hadoop.hbase.quotas.TestQuotaThrottle.testUserNamespaceReadAndWriteThrottle(TestQuotaThrottle.java:276) at org.apache.hadoop.hbase.replication.TestMasterReplication.testHFileReplicationForConfiguredTableCfs(TestMasterReplication.java:347) at org.apache.hadoop.hbase.regionserver.TestHRegion.testFlushCacheWhileScanning(TestHRegion.java:3824) Test results: https://builds.apache.org/job/PreCommit-PHOENIX-Build/645//testReport/ Javadoc warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/645//artifact/patchprocess/patchJavadocWarnings.txt Console output: https://builds.apache.org/job/PreCommit-PHOENIX-Build/645//console This message is automatically generated. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was >
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15609909#comment-15609909 ] James Taylor commented on PHOENIX-476: -- Patch looks good, [~kliew]. I made the following few minor changes and will commit later today: - Added test and code to deal with type mismatch and size compatibility check at compile time. In general, always best to do these checks as early as possible. Checking when the DefaultValueExpression is created is too late, as then the table would have an invalid default expression. - Moved the check for isRowTimestamp to CreateTableCompiler to avoid necessity of funky isIndexColumn check. - Moved negative tests to QueryCompilerTest as compile time tests don't need to spin up a mini cluster. - Removed throw of SQLException from PTable.newKey() as it's never actually thrown. - Removed default expression if it evaluates to null as that doesn't add any value. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.12.patch, PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15602775#comment-15602775 ] James Taylor commented on PHOENIX-476: -- Excellent, [~kliew]. Is your latest patch good-to-go, or is there still more to do? Thanks! > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.2.patch, PHOENIX-476.3.patch, PHOENIX-476.4.patch, > PHOENIX-476.5.patch, PHOENIX-476.6.patch, PHOENIX-476.7.patch, > PHOENIX-476.8.patch, PHOENIX-476.9.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15601143#comment-15601143 ] Kevin Liew commented on PHOENIX-476: Thanks [~jamestaylor], that fixed it. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.2.patch, PHOENIX-476.3.patch, PHOENIX-476.4.patch, > PHOENIX-476.5.patch, PHOENIX-476.6.patch, PHOENIX-476.7.patch, > PHOENIX-476.8.patch, PHOENIX-476.9.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15601001#comment-15601001 ] Hadoop QA commented on PHOENIX-476: --- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12834871/PHOENIX-476.11.patch against master branch at commit 8f5c8cfbd1c1f7894e4cc88e2257d245d1d8c3bf. ATTACHMENT ID: 12834871 {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 3 new or modified tests. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:red}-1 javadoc{color}. The javadoc tool appears to have generated 43 warning messages. {color:red}-1 release audit{color}. The applied patch generated 1 release audit warnings (more than the master's current 0 warnings). {color:red}-1 lineLengths{color}. The patch introduces the following lines longer than 100: +conn.createStatement().executeQuery("SELECT c1 FROM " + sharedTable2 + " WHERE c1 = 10"); +rs = conn.createStatement().executeQuery("SELECT c4 FROM " + sharedTable2 + " WHERE c4 = 'ABCD'"); +: c=column_name dt=identifier (LPAREN l=NUMBER (COMMA s=NUMBER)? RPAREN)? ar=ARRAY? (lsq=LSQUARE (a=NUMBER)? RSQUARE)? (nn=NOT? n=NULL)? (DEFAULT df=expression)? (pk=PRIMARY KEY (order=ASC|order=DESC)? rr=ROW_TIMESTAMP?)? +protected RegionScanner doPostScannerOpen(final ObserverContext c, final Scan scan, final RegionScanner s) throws IOException, SQLException { +CANNOT_CREATE_DEFAULT_ROWTIMESTAMP(1032, "42Y90", "Cannot create ROW_TIMESTAMP column with a default value."), +public ColumnDef columnDef(ColumnName columnDefName, String sqlTypeName, boolean isArray, Integer arrSize, Boolean isNull, Integer maxLength, Integer scale, boolean isPK, +public Expression newColumnExpression(boolean schemaNameCaseSensitive, boolean colNameCaseSensitive) throws SQLException { +columnDefs.add(FACTORY.columnDef(colName, col.getDataType().getSqlTypeName(), col.isNullable(), col.getMaxLength(), col.getScale(), false, col.getSortOrder(), col.getExpressionStr(), col.isRowTimestamp())); +PColumn newColumn = new PColumnImpl(oldColumn.getName(), oldColumn.getFamilyName(), oldColumn.getDataType(), oldColumn.getMaxLength(), oldColumn.getScale(), oldColumn.isNullable(), i-1+positionOffset, oldColumn.getSortOrder(), oldColumn.getArraySize(), oldColumn.getViewConstant(), oldColumn.isViewReferenced(), oldColumn.getExpressionStr(), oldColumn.isRowTimestamp(), oldColumn.isDynamic()); + public void coerceBytes(ImmutableBytesWritable ptr, Object o, PDataType actualType, Integer actualMaxLength, {color:green}+1 core tests{color}. The patch passed unit tests in . Test results: https://builds.apache.org/job/PreCommit-PHOENIX-Build/636//testReport/ Release audit warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/636//artifact/patchprocess/patchReleaseAuditWarnings.txt Javadoc warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/636//artifact/patchprocess/patchJavadocWarnings.txt Console output: https://builds.apache.org/job/PreCommit-PHOENIX-Build/636//console This message is automatically generated. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.2.patch, PHOENIX-476.3.patch, PHOENIX-476.4.patch, > PHOENIX-476.5.patch, PHOENIX-476.6.patch, PHOENIX-476.7.patch, > PHOENIX-476.8.patch, PHOENIX-476.9.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15600941#comment-15600941 ] Hadoop QA commented on PHOENIX-476: --- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12834870/PHOENIX-476.10.patch against master branch at commit 8f5c8cfbd1c1f7894e4cc88e2257d245d1d8c3bf. ATTACHMENT ID: 12834870 {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 3 new or modified tests. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:red}-1 javadoc{color}. The javadoc tool appears to have generated 43 warning messages. {color:red}-1 release audit{color}. The applied patch generated 1 release audit warnings (more than the master's current 0 warnings). {color:red}-1 lineLengths{color}. The patch introduces the following lines longer than 100: +conn.createStatement().executeQuery("SELECT c1 FROM " + sharedTable2 + " WHERE c1 = 10"); +rs = conn.createStatement().executeQuery("SELECT c4 FROM " + sharedTable2 + " WHERE c4 = 'ABCD'"); +: c=column_name dt=identifier (LPAREN l=NUMBER (COMMA s=NUMBER)? RPAREN)? ar=ARRAY? (lsq=LSQUARE (a=NUMBER)? RSQUARE)? (nn=NOT? n=NULL)? (DEFAULT df=expression)? (pk=PRIMARY KEY (order=ASC|order=DESC)? rr=ROW_TIMESTAMP?)? +protected RegionScanner doPostScannerOpen(final ObserverContext c, final Scan scan, final RegionScanner s) throws IOException, SQLException { +CANNOT_CREATE_DEFAULT_ROWTIMESTAMP(1032, "42Y90", "Cannot create ROW_TIMESTAMP column with a default value."), +public ColumnDef columnDef(ColumnName columnDefName, String sqlTypeName, boolean isArray, Integer arrSize, Boolean isNull, Integer maxLength, Integer scale, boolean isPK, +public Expression newColumnExpression(boolean schemaNameCaseSensitive, boolean colNameCaseSensitive) throws SQLException { +columnDefs.add(FACTORY.columnDef(colName, col.getDataType().getSqlTypeName(), col.isNullable(), col.getMaxLength(), col.getScale(), false, col.getSortOrder(), col.getExpressionStr(), col.isRowTimestamp())); +PColumn newColumn = new PColumnImpl(oldColumn.getName(), oldColumn.getFamilyName(), oldColumn.getDataType(), oldColumn.getMaxLength(), oldColumn.getScale(), oldColumn.isNullable(), i-1+positionOffset, oldColumn.getSortOrder(), oldColumn.getArraySize(), oldColumn.getViewConstant(), oldColumn.isViewReferenced(), oldColumn.getExpressionStr(), oldColumn.isRowTimestamp(), oldColumn.isDynamic()); + public void coerceBytes(ImmutableBytesWritable ptr, Object o, PDataType actualType, Integer actualMaxLength, {color:red}-1 core tests{color}. The patch failed these unit tests: ./phoenix-core/target/failsafe-reports/TEST-org.apache.phoenix.end2end.AlterTableIT ./phoenix-core/target/failsafe-reports/TEST-[ViewIndexIT_isNamespaceMapped=false] Test results: https://builds.apache.org/job/PreCommit-PHOENIX-Build/635//testReport/ Release audit warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/635//artifact/patchprocess/patchReleaseAuditWarnings.txt Javadoc warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/635//artifact/patchprocess/patchJavadocWarnings.txt Console output: https://builds.apache.org/job/PreCommit-PHOENIX-Build/635//console This message is automatically generated. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.10.patch, PHOENIX-476.11.patch, > PHOENIX-476.2.patch, PHOENIX-476.3.patch, PHOENIX-476.4.patch, > PHOENIX-476.5.patch, PHOENIX-476.6.patch, PHOENIX-476.7.patch, > PHOENIX-476.8.patch, PHOENIX-476.9.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null.
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15600521#comment-15600521 ] James Taylor commented on PHOENIX-476: -- Patch is looking good. For checking isRowTimestamp set in PrimaryKeyConstraint, feel free to add check here in MetaDataClient: {code} private PColumn newColumn(int position, ColumnDef def, PrimaryKeyConstraint pkConstraint, String defaultColumnFamily, boolean addingToPK) throws SQLException { try { ColumnName columnDefName = def.getColumnDefName(); SortOrder sortOrder = def.getSortOrder(); boolean isPK = def.isPK(); boolean isRowTimestamp = def.isRowTimestamp(); if (pkConstraint != null) { PairpkSortOrder = pkConstraint.getColumnWithSortOrder(columnDefName); if (pkSortOrder != null) { isPK = true; sortOrder = pkSortOrder.getSecond(); isRowTimestamp = pkConstraint.isColumnRowTimestamp(columnDefName); } } {code} For covered columns, make sure the default value gets passed into the ColumnDef being created here in MetaDataClient.createIndex() call. That should be all that's required: {code} // Last all the included columns (minus any PK columns) for (ColumnName colName : includedColumns) { PColumn col = resolver.resolveColumn(null, colName.getFamilyName(), colName.getColumnName()).getColumn(); colName = ColumnName.caseSensitiveColumnName(IndexUtil.getIndexColumnName(col)); // Check for duplicates between indexed and included columns if (indexedColumnNames.contains(colName)) { throw new SQLExceptionInfo.Builder(SQLExceptionCode.COLUMN_EXIST_IN_DEF).build().buildException(); } if (!SchemaUtil.isPKColumn(col) && col.getViewConstant() == null) { // Need to re-create ColumnName, since the above one won't have the column family name colName = ColumnName.caseSensitiveColumnName(isLocalIndex?IndexUtil.getLocalIndexColumnFamily(col.getFamilyName().getString()):col.getFamilyName().getString(), IndexUtil.getIndexColumnName(col)); columnDefs.add(FACTORY.columnDef(colName, col.getDataType().getSqlTypeName(), col.isNullable(), col.getMaxLength(), col.getScale(), false, col.getSortOrder(), null, col.isRowTimestamp())); } } {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15600113#comment-15600113 ] Hadoop QA commented on PHOENIX-476: --- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12834851/PHOENIX-476.9.patch against master branch at commit 8f5c8cfbd1c1f7894e4cc88e2257d245d1d8c3bf. ATTACHMENT ID: 12834851 {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 3 new or modified tests. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:red}-1 javadoc{color}. The javadoc tool appears to have generated 43 warning messages. {color:red}-1 release audit{color}. The applied patch generated 1 release audit warnings (more than the master's current 0 warnings). {color:red}-1 lineLengths{color}. The patch introduces the following lines longer than 100: + assertEquals(SQLExceptionCode.CANNOT_CREATE_DEFAULT_ROWTIMESTAMP.getErrorCode(), e.getErrorCode()); +: c=column_name dt=identifier (LPAREN l=NUMBER (COMMA s=NUMBER)? RPAREN)? ar=ARRAY? (lsq=LSQUARE (a=NUMBER)? RSQUARE)? (nn=NOT? n=NULL)? (DEFAULT df=expression)? (pk=PRIMARY KEY (order=ASC|order=DESC)? rr=ROW_TIMESTAMP?)? +if (columnDef.getColumnDefName().getFamilyName() != null && columnDef.getColumnDefName().getFamilyName().contains(QueryConstants.LOCAL_INDEX_COLUMN_FAMILY_PREFIX)) { +throw new SQLExceptionInfo.Builder(SQLExceptionCode.CANNOT_CREATE_DEFAULT_ROWTIMESTAMP) + .setColumnName(columnDef.getColumnDefName().getColumnName()).build().buildException(); +ParseNode defaultParseNode = new SQLParser(columnDef.getExpression()).parseExpression(); +if (!defaultParseNode.isStateless() || defaultExpression.getDeterminism() != Determinism.ALWAYS) { + .setColumnName(columnDef.getColumnDefName().getColumnName()).build().buildException(); +protected RegionScanner doPostScannerOpen(final ObserverContext c, final Scan scan, final RegionScanner s) throws IOException, SQLException { +CANNOT_CREATE_DEFAULT_ROWTIMESTAMP(1032, "42Y90", "Cannot create ROW_TIMESTAMP column with a default value."), {color:red}-1 core tests{color}. The patch failed these unit tests: ./phoenix-core/target/failsafe-reports/TEST-org.apache.phoenix.end2end.AlterTableIT Test results: https://builds.apache.org/job/PreCommit-PHOENIX-Build/634//testReport/ Release audit warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/634//artifact/patchprocess/patchReleaseAuditWarnings.txt Javadoc warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/634//artifact/patchprocess/patchJavadocWarnings.txt Console output: https://builds.apache.org/job/PreCommit-PHOENIX-Build/634//console This message is automatically generated. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.9.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598934#comment-15598934 ] Kevin Liew commented on PHOENIX-476: I filed PHOENIX-3397 for that. Also, PHOENIX-3398 regarding `isRowTimestamp` not being set when we set ROW_TIMESTAMP in the CONSTRAINTS field, which should be possible according to the website. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598665#comment-15598665 ] James Taylor commented on PHOENIX-476: -- Still sounds like a bug lurking here, but it doesn't need to be fixed for this JIRA. But let's file one with the above info so we don't lose track of it. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598651#comment-15598651 ] Kevin Liew commented on PHOENIX-476: I got it working. With the recent changes, we access the context again during the UPSERT so we can compile arbitrary isConstant expressions. I just needed to revert the older changes where we were compiling the expression to a literal in CreateTableCompiler and storing the expression string obtained from that. For scalar date-time, the expression was untouched. For arrays, it was being modified below: {code:java} public Expression visitLeave(ArrayConstructorNode node, List children) throws SQLException { ... ArrayConstructorExpression arrayExpression = new ArrayConstructorExpression(children, arrayElemDataType, rowKeyOrderOptimizable); if (ExpressionUtil.isConstant(arrayExpression)) { for (int i = 0; i < children.size(); i++) { Expression child = children.get(i); child.evaluate(null, ptr); Object value = null; if (child.getDataType() == null) { value = arrayElemDataType.toObject(ptr, theArrayElemDataType, child.getSortOrder()); } else { value = arrayElemDataType.toObject(ptr, child.getDataType(), child.getSortOrder()); } elements[i] = LiteralExpression.newConstant(value, theArrayElemDataType, child.getDeterminism()).getValue(); } Object value = PArrayDataType.instantiatePhoenixArray(arrayElemDataType, elements); return LiteralExpression.newConstant(value, PDataType.fromTypeId(arrayElemDataType.getSqlType() + PDataType.ARRAY_TYPE_BASE), null, null, arrayExpression.getSortOrder(), Determinism.ALWAYS, rowKeyOrderOptimizable); } {code} Its just testing and cleanup left now. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598637#comment-15598637 ] James Taylor commented on PHOENIX-476: -- Good find. Please file a separate JIRA for that and don't worry about it for now (i.e. Don't use date/time types in arrays for your tests). > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598382#comment-15598382 ] Kevin Liew commented on PHOENIX-476: We compile the string to an expression contained in a ColumnDef. For arrays: the date-time is represented as a string (without the type specifier) in the new expression string. When we call `parseExpression`, it only sees the string without the type information. {code:java} ExpressionCompiler compiler = new ExpressionCompiler(context); ParseNode defaultParseNode = new SQLParser(column.getExpressionStr()).parseExpression(); Expression defaultExpression = defaultParseNode.accept(compiler); {code} For scalar columns, the type specifier is not removed after compiling an Expression (ie. "DATE '1900-10-01 14:03:22.559'"). I don't think it is expected for the type specifier to be removed after compiling array values. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598342#comment-15598342 ] James Taylor commented on PHOENIX-476: -- Looks correct - you can verify by using a SELECT statement with the same literal expression. You're still using parseExpression(), right? Might be a bug. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598326#comment-15598326 ] Kevin Liew commented on PHOENIX-476: The byte coercion fix for data types also fixed byte coercion for arrays. Right now I'm running into a problem with default values for date-time types. "time '1900-10-01 14:03:22.559'" worked for non array types. I tried a few different variations of: {code:sql} tim TIME[5] DEFAULT ARRAY[time '1900-10-01 14:03:22.559', time '1901-10-01 14:03:22.559'] {code} {code} org.apache.phoenix.schema.TypeMismatchException: ERROR 203 (22005): Type mismatch. ERROR 203 (22005): Type mismatch. ERROR 203 (22005): Type mismatch. ERROR 203 (22005): Type mismatch. VARCHAR cannot be coerced to TIME at org.apache.phoenix.exception.SQLExceptionCode$1.newException(SQLExceptionCode.java:74) at org.apache.phoenix.exception.SQLExceptionInfo.buildException(SQLExceptionInfo.java:145) at org.apache.phoenix.util.ServerUtil.parseRemoteException(ServerUtil.java:129) at org.apache.phoenix.util.ServerUtil.parseServerExceptionOrNull(ServerUtil.java:118) at org.apache.phoenix.util.ServerUtil.parseServerException(ServerUtil.java:107) at org.apache.phoenix.iterate.BaseResultIterators.getIterators(BaseResultIterators.java:751) at org.apache.phoenix.iterate.BaseResultIterators.getIterators(BaseResultIterators.java:695) at org.apache.phoenix.iterate.ConcatResultIterator.getIterators(ConcatResultIterator.java:50) at org.apache.phoenix.iterate.ConcatResultIterator.currentIterator(ConcatResultIterator.java:97) at org.apache.phoenix.iterate.ConcatResultIterator.next(ConcatResultIterator.java:117) at org.apache.phoenix.jdbc.PhoenixResultSet.next(PhoenixResultSet.java:778) at org.apache.phoenix.end2end.DefaultColumnValueIT.testDefaultArrays(DefaultColumnValueIT.java:732) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:498) at org.junit.runners.model.FrameworkMethod$1.runReflectiveCall(FrameworkMethod.java:50) at org.junit.internal.runners.model.ReflectiveCallable.run(ReflectiveCallable.java:12) at org.junit.runners.model.FrameworkMethod.invokeExplosively(FrameworkMethod.java:47) at org.junit.internal.runners.statements.InvokeMethod.evaluate(InvokeMethod.java:17) at org.junit.internal.runners.statements.RunAfters.evaluate(RunAfters.java:27) at org.junit.runners.ParentRunner.runLeaf(ParentRunner.java:325) at org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:78) at org.junit.runners.BlockJUnit4ClassRunner.runChild(BlockJUnit4ClassRunner.java:57) at org.junit.runners.ParentRunner$3.run(ParentRunner.java:290) at org.junit.runners.ParentRunner$1.schedule(ParentRunner.java:71) at org.junit.runners.ParentRunner.runChildren(ParentRunner.java:288) at org.junit.runners.ParentRunner.access$000(ParentRunner.java:58) at org.junit.runners.ParentRunner$2.evaluate(ParentRunner.java:268) at org.junit.internal.runners.statements.RunBefores.evaluate(RunBefores.java:26) at org.junit.internal.runners.statements.RunAfters.evaluate(RunAfters.java:27) at org.junit.rules.ExternalResource$1.evaluate(ExternalResource.java:48) at org.junit.rules.RunRules.evaluate(RunRules.java:20) at org.junit.runners.ParentRunner.run(ParentRunner.java:363) at org.junit.runner.JUnitCore.run(JUnitCore.java:137) at com.intellij.junit4.JUnit4IdeaTestRunner.startRunnerWithArgs(JUnit4IdeaTestRunner.java:117) at com.intellij.junit4.JUnit4IdeaTestRunner.startRunnerWithArgs(JUnit4IdeaTestRunner.java:42) at com.intellij.rt.execution.junit.JUnitStarter.prepareStreamsAndStart(JUnitStarter.java:262) at com.intellij.rt.execution.junit.JUnitStarter.main(JUnitStarter.java:84) at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62) at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43) at java.lang.reflect.Method.invoke(Method.java:498) at com.intellij.rt.execution.application.AppMain.main(AppMain.java:147) {code} I'll try alternate syntax, but was the above syntactically correct? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598316#comment-15598316 ] Hadoop QA commented on PHOENIX-476: --- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12834822/PHOENIX-476.8.patch against master branch at commit 8f5c8cfbd1c1f7894e4cc88e2257d245d1d8c3bf. ATTACHMENT ID: 12834822 {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 3 new or modified tests. {color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings. {color:red}-1 javadoc{color}. The javadoc tool appears to have generated 43 warning messages. {color:red}-1 release audit{color}. The applied patch generated 1 release audit warnings (more than the master's current 0 warnings). {color:red}-1 lineLengths{color}. The patch introduces the following lines longer than 100: +ResultSet rs = conn.createStatement().executeQuery("SELECT " + projection + " FROM T_DEFAULT WHERE pk1 = 1"); +rs = conn.createStatement().executeQuery("SELECT " + projection + " FROM T_DEFAULT WHERE pk1 = 11"); +ResultSet rs = conn.createStatement().executeQuery("SELECT " + projection + " FROM T_DEFAULT WHERE pk1 = 1"); +rs = conn.createStatement().executeQuery("SELECT " + projection + " FROM T_DEFAULT WHERE pk1 = 11"); +rs = conn.createStatement().executeQuery("SELECT " + projection + " FROM T_DEFAULT WHERE pk1 = 1"); +rs = conn.createStatement().executeQuery("SELECT " + projection + " FROM T_DEFAULT WHERE pk1 = 11"); +ResultSet rs = conn.createStatement().executeQuery("SELECT * FROM T_DEFAULT2 WHERE pk1 = 1"); +ResultSet rs = conn.createStatement().executeQuery("SELECT * FROM T_DEFAULT2 WHERE pk1 = 1"); +ResultSet rs = conn.createStatement().executeQuery("SELECT * FROM " + table + " WHERE pk = 1"); +ResultSet rs = conn.createStatement().executeQuery("SELECT * FROM " + table + " WHERE pk = 1"); {color:red}-1 core tests{color}. The patch failed these unit tests: ./phoenix-core/target/failsafe-reports/TEST-org.apache.phoenix.end2end.DefaultColumnValueIT ./phoenix-core/target/failsafe-reports/TEST-org.apache.phoenix.end2end.CreateSchemaIT Test results: https://builds.apache.org/job/PreCommit-PHOENIX-Build/633//testReport/ Release audit warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/633//artifact/patchprocess/patchReleaseAuditWarnings.txt Javadoc warnings: https://builds.apache.org/job/PreCommit-PHOENIX-Build/633//artifact/patchprocess/patchJavadocWarnings.txt Console output: https://builds.apache.org/job/PreCommit-PHOENIX-Build/633//console This message is automatically generated. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598313#comment-15598313 ] James Taylor commented on PHOENIX-476: -- Is anything special required for array support? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598280#comment-15598280 ] Kevin Liew commented on PHOENIX-476: Thanks James, that works. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598264#comment-15598264 ] James Taylor commented on PHOENIX-476: -- Scratch that - it's not going to work to have a null PhoenixConnection. Do this instead inside of PTableImpl: {code} String url = PhoenixRuntime.JDBC_PROTOCOL + PhoenixRuntime.JDBC_PROTOCOL_SEPARATOR + PhoenixRuntime.CONNECTIONLESS PhoenixConnection conn = DriverManager.getConnection(url).unwrap(PhoenixConnection.class); StatementContext context = new StatementContext(new PhoenixStatement(conn)); {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598256#comment-15598256 ] James Taylor commented on PHOENIX-476: -- I see - yes that'll still be a problem. Try this to create a StatementContext within the PTableImpl.newKey() instead of passing it in: {code} StatementContext context = new StatementContext(new PhoenixStatement(null)); {code} I think you might be able to get away with having a null Connection when you're only parsing an expression that is a literal. If that works, you can still with your last patch, but without passing through a StatementContext to the PTableImpl method. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598254#comment-15598254 ] Hadoop QA commented on PHOENIX-476: --- {color:red}-1 overall{color}. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12834821/PHOENIX-476.8.patch against master branch at commit 8f5c8cfbd1c1f7894e4cc88e2257d245d1d8c3bf. ATTACHMENT ID: 12834821 {color:green}+1 @author{color}. The patch does not contain any @author tags. {color:green}+1 tests included{color}. The patch appears to include 69 new or modified tests. {color:red}-1 patch{color}. The patch command could not apply the patch. Console output: https://builds.apache.org/job/PreCommit-PHOENIX-Build/632//console This message is automatically generated. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598242#comment-15598242 ] Kevin Liew commented on PHOENIX-476: Oops, sorry I had to reset my environment so the patch is messy due to IDE changes. I will clean up the patch and attach again. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.8.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598232#comment-15598232 ] James Taylor commented on PHOENIX-476: -- So for this to work, in CreateTableCompiler, after you evaluate the DEFAULT expression, you'll need to replace the existing ColumnDef with a new one that you'll instantiate using the same values except for the expressionStr which you'll replace with the evaluatedDefaultExpression.toString() value. {code} ColumnDef(ColumnName columnDefName, String sqlTypeName, boolean isArray, Integer arrSize, Boolean isNull, Integer maxLength, Integer scale, boolean isPK, SortOrder sortOrder, String expressionStr, boolean isRowTimestamp) { {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598230#comment-15598230 ] Kevin Liew commented on PHOENIX-476: That's what I did, but the literal `-100` fails to parse as a literal. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598225#comment-15598225 ] James Taylor commented on PHOENIX-476: -- No parser changes are required: -In CreateTableCompiler, when you compile the defaultExpression, you'll end up with a LiteralExpression (i.e. DEFAULT -2 + 7 would be turned into a LiteralExpression of 5). Create a new CreateTableStatement copy constructor that passes in a createTableStatement and an expressionStr. Use this to create a new CreateTableStatement using defaultExpression.toString() to form the literal expression string which you'll pass through to the MetaDataClient to guarantee that the default expression is always a literal. - By evaluating the default expression at CREATE TABLE time, and storing the evaluated expression string value, you'll know you have literal. When reading and parsing the default value back in PTableImpl, you can assume it's already a literal and parse it as you were doing in your previous patch: {code} +LiteralParseNode defaultParseNode = new SQLParser(column.getExpressionStr()).parseLiteral(); +LiteralExpression defaultLiteral = LiteralExpression.newConstant(defaultParseNode.getValue(), column.getDataType(), column.getMaxLength(), column.getScale(), column.getSortOrder(), Determinism.ALWAYS); +ImmutableBytesWritable valuePtr = new ImmutableBytesWritable(); +defaultLiteral.evaluate(null, valuePtr); +byteValue = ByteUtil.copyKeyBytesIfNecessary(valuePtr); {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15598196#comment-15598196 ] Kevin Liew commented on PHOENIX-476: One possible solution is to create a second parser class specifically for literals, which would give us the freedom to define any tokens without affecting expression parsing. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15597332#comment-15597332 ] Kevin Liew commented on PHOENIX-476: Hi [~jamestaylor], I made the suggested changes to recreate the create-table statement with the literal value string that we get after compiling the statement. Without a StatementContext, that takes us back to the issue with parsing negative literals at `new SQLParser(column.getExpressionStr()).parseLiteral()`. I tried defining a new parse function without adding new tokens so that other functions won't be affected but the parser throws an exception. {code} // Get a string, integer, double, date, boolean, or NULL value. // Able to parse negative numbers. any_literal returns [LiteralParseNode ret] : s=STRING_LITERAL { ret = factory.literal(s.getText()); } | n=NUMBER { ret = factory.wholeNumber(n.getText()); } | nn=MINUS DIGIT+ { ret = factory.wholeNumber(nn.getText()); } | d=DECIMAL { ret = factory.realNumber(d.getText()); } | nd=MINUS DIGIT+ DOT POSINTEGER { ret = factory.realNumber(nd.getText()); } | dbl=DOUBLE { ret = factory.literal(Double.valueOf(dbl.getText())); } | ndbl='.' POSINTEGER Exponent | MINUS DIGIT+ DOT Exponent | MINUS DIGIT+ (DOT (POSINTEGER (Exponent)?)? | Exponent) { ret = factory.literal(Double.valueOf(ndbl.getText())); } | NULL {ret = factory.literal(null);} | TRUE {ret = factory.literal(Boolean.TRUE);} | FALSE {ret = factory.literal(Boolean.FALSE);} | dt=identifier t=STRING_LITERAL { try { ret = factory.literal(t.getText(), dt); } catch (SQLException e) { throw new RuntimeException(e); } } ; {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15590451#comment-15590451 ] Kevin Liew commented on PHOENIX-476: Sure thing. I'll get started on the revision and cleanup. I'll take a closer look at array support for a better estimate but it might be a weekend-task. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15590310#comment-15590310 ] James Taylor commented on PHOENIX-476: -- Ping [~kliew] - trying to do some release planning - do you think you'd have some spare cycles to finish this up soon? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15566032#comment-15566032 ] James Taylor commented on PHOENIX-476: -- One more error check too: - In CreateTableCompiler, if a columnDef defines a default value, check that columnDef.isRowTimestamp() is false. It wouldn't make sense to define a default value for the ROW_TIMESTAMP column. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15564469#comment-15564469 ] James Taylor commented on PHOENIX-476: -- Couple more important tests around secondary indexes: - create an index with a covered column that has a default value and make sure that the default value takes affect. - create an index on a column with a default and make sure the index rows are correct. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15564303#comment-15564303 ] James Taylor commented on PHOENIX-476: -- Let's get array support in and the other revisions I mentioned, but do the other work in follow up JIRAs. It looks like this can make 4.9.0 which is great! > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15564286#comment-15564286 ] Kevin Liew commented on PHOENIX-476: Oh, right - I was thinking about backward compatibility from the server pov but the user pov sees broken compatibility if they have to use a new name for the default schema. The tests are all passing locally and I think [~hadoopqa] will post results in a little bit. I will revise based on your feedback and also take another look at support for arrays. I made some changes to type coercion in DefaultValueExpression for the data type test cases and this may resolve the coercion issues I was seeing with arrays. If there is no urgency in getting this committed, we could commit array support with this patch. We could also branch array support into a later JIRA, as well as stateful expressions and additional DDL support. And support for `DEFAULT` in `UPSERT` statements as Julian had mentioned. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15564198#comment-15564198 ] James Taylor commented on PHOENIX-476: -- Thanks for the quick turnaround, [~kliew]. I think we're almost there. Do the tests pass now? A few changes are needed, though. - To maintain backwards compatibility (as your change would break usages of USE SCHEMA DEFAULT), you can make this change to PhoenixSQL.g: {code} use_schema_node returns [UseSchemaStatement ret] : USE (DEFAULT | s=identifier) {ret = factory.useSchema(s); } ; {code} And then in UseSchemaStatement: {code} public class UseSchemaStatement extends MutableStatement { private final String schemaName; public UseSchemaStatement(String schemaName) { this.schemaName = schemaName == null ? StringUtil.EMPTY_STRING : schemaName; } {code} And then finally in SchemaUtil, make this change: {code} public MutationState useSchema(UseSchemaStatement useSchemaStatement) throws SQLException { // As we allow default namespace mapped to empty schema, so this is to reset schema in connection if (useSchemaStatement.getSchemaName().equals(StringUtil.EMPTY_STRING) { connection.setSchema(null); } else { {code} - Using null in UngroupedAggregateRegionObserver will be a problem. Add a test for a table with a default value for one of the PK columns and do an UPSERT SELECT into it. {code} diff --git a/phoenix-core/src/main/java/org/apache/phoenix/coprocessor/UngroupedAggregateRegionObserver.java b/phoenix-core/src/main/java/org/apache/phoenix/coprocessor/UngroupedAggregateRegionObserver.java index d8d313d..c30a44b 100644 --- a/phoenix-core/src/main/java/org/apache/phoenix/coprocessor/UngroupedAggregateRegionObserver.java +++ b/phoenix-core/src/main/java/org/apache/phoenix/coprocessor/UngroupedAggregateRegionObserver.java @@ -435,7 +435,7 @@ public class UngroupedAggregateRegionObserver extends BaseScannerRegionObserver } values[i] = ptr.copyBytes(); } -writeToTable.newKey(ptr, values); +writeToTable.newKey(null, ptr, values); {code} Rather than requiring a StatementContext, let's do the following: - In CreateTableCompiler, when you compile the defaultExpression, you'll end up with a LiteralExpression (i.e. DEFAULT 2 + 7 would be turned into a LiteralExpression of 9). Create a new CreateTableStatement copy constructor that passes in a createTableStatement and an expressionStr. Use this to create a new CreateTableStatement using defaultExpression.toString() to form the literal expression string which you'll pass through to the MetaDataClient to guarantee that the default expression is always a literal. - Then modify the usage of the default value back to the way you had it where you didn't need a StatementContext since you'll know you have a literal. We'll may have to tweak this once we want to support NEXT VALUE FOR and CURRENT_DATE expressions (or we may not need to as these values are generated on the client and turned into literals anyway). But for now, this will work fine. - Also, instead of deriving DefaultColumnValueIT from BaseClientManagedTimeIT, derive it from ParallelStatsDisabledIT. Then just change hardcoded references of T_DEFAULT to a new fullTableName member variable which you set in an @Before method using the static generateUniqueName() method. This allows your tests to run in parallel (since the table names are independent) and enables our tests to run much faster. The tests in BaseClientManagedTimeIT are the only ones waiting to be converted like this as well, so it's essentially deprecated. You also don't need to use nextTimestamp() and open new connections every time - just use the same connection throughout your test. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Fix For: 4.9.0 > > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.7.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15562842#comment-15562842 ] James Taylor commented on PHOENIX-476: -- Thanks, [~kliew]. Looking very good. Couple of comments: - Why do you say that without caching expression we can only compile literals? The SQLParser.parseExpression() call will happily parse any expression and the ExpressionCompiler will translate from the ParseNode to an Expression. The caching of expressions was just an optimization, not a requirement. Also, we don't parse negative literals because it leads to problems with certain expressions. For example, QueryParserTest.testSubtractionInSelect doesn't pass with your patch applied. - Minor, but I think the change to this file can be reverted: {code} diff --git a/phoenix-core/src/main/java/org/apache/phoenix/schema/MetaDataClient.java b/phoenix-core/src/main/java/org/apache/phoenix/schema/MetaDataClient.java index 1e635d8..f617f92 100644 --- a/phoenix-core/src/main/java/org/apache/phoenix/schema/MetaDataClient.java +++ b/phoenix-core/src/main/java/org/apache/phoenix/schema/MetaDataClient.java @@ -431,7 +431,7 @@ public class MetaDataClient { /** * Update the cache with the latest as of the connection scn. - * @param functioNames + * @param functionNames * @return the timestamp from the server, negative if the function was added to the cache and positive otherwise * @throws SQLException */ {code} If you can undo the changes for the parser for negative literals, use SQLParser.parseExpression() instead of parseLiteral(), add a few of your test back that use expressions instead of literals, and ensure that all unit tests pass (i.e. run {{mvn verify}}), then I think we'll be good to go. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.6.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15560493#comment-15560493 ] James Taylor commented on PHOENIX-476: -- Looking really good, [~kliew]. Thanks for the updates! Here's some feedback: - Don't bother caching the default Expression. It's not worth the trouble as the parsing cost is minuscule. Your technique will break down if a column is dropped as the position changes in that case. This could lead to a default value of what was in a lower position to be used mistakenly. Just not caching it is fine. If it ever shows up as a perf issue, we can deal with it then. - Instead of forcing STORE_NULLS to be true for a table with default values (and I should have thought about this before), just tweak the code in PRowImpl.setValue() based on if the PColumn has a default value. Something like this: {code} @Override public void setValue(PColumn column, byte[] byteValue) { deleteRow = null; byte[] family = column.getFamilyName().getBytes(); byte[] qualifier = column.getName().getBytes(); PDataType type = column.getDataType(); // Check null, since some types have no byte representation for null boolean isNull = type.isNull(byteValue); if (isNull && !column.isNullable()) { throw new ConstraintViolationException(name.getString() + "." + column.getName().getString() + " may not be null"); } else if (isNull && PTableImpl.this.isImmutableRows() && column.getExpressionStr() == null) { // Need to preserve explicit nulls for immutable tables otherwise default value would be used removeIfPresent(setValues, family, qualifier); removeIfPresent(unsetValues, family, qualifier); } else if (isNull && !getStoreNulls() && && column.getExpressionStr() == null) { {code} - Add some test cases where the table is immutable (i.e. set IMMUTABLE_ROWS=true on DDL statement) - Add a test for the case of a default value being set on a PK column in the middle. For example: PK1, PK2 DEFAULT 'foo', PK3. - Also, have a test that uses different types for default values (i.e. VARCHAR, DECIMAL, TINYINT, and ARRAY would be interesting too). - Make sure you've imported all of our default settings in terms of keeping imports explicitly rather than wildcarding, import order, etc. For example, I noticed that this was incorrectly change by your IDE: {code} -import static org.junit.Assert.assertEquals; -import static org.junit.Assert.assertNotEquals; -import static org.junit.Assert.assertNotNull; -import static org.junit.Assert.assertTrue; -import static org.junit.Assert.fail; - -import java.sql.Connection; -import java.sql.DriverManager; -import java.sql.SQLException; -import java.sql.Statement; +import static org.junit.Assert.*; {code} See https://phoenix.apache.org/contributing.html#Code_conventions > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.4.patch, PHOENIX-476.5.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15558377#comment-15558377 ] James Taylor commented on PHOENIX-476: -- Thanks for the patch, [~kliew]. Here's some feedback: - Let's revert this change to ExpressionUtil as it's creating too many diffs in other places. Plus, you'll only use the logic to check that an expression is a constant in one place: {code} - public static boolean isConstant(Expression expression) { +public static boolean isConstant(Expression expression) { +return (expression.isStateless() && expression.getDeterminism() == Determinism.ALWAYS); +} + + public static boolean isPureExpression(Expression expression) { {code} - This change in ExpressionCompiler can be much simpler. No need to recreate a ColumnDef. Also, PColumn needs to stay immutable, so no setters there. Instead, just parse the expressionStr using {{expressionNode = new SQLParser(expressionStr).parseExpression()}} and call expressionNode.accept(this) to get the Expression here (and minor nit, but just check that {{column.getExpressionStr() != null}}. Shouldn't be any need to check if empty too): {code} +if (!SchemaUtil.isPKColumn(column)) { +if (column.getExpressionStr() != null && !column.getExpressionStr().isEmpty() && column.getDefaultExpression() == null) { +// We have the default expression but it has not been compiled yet +ExpressionCompiler compiler = new ExpressionCompiler(context); +ColumnDef columnDef = new ParseNodeFactory().columnDef( + ColumnName.caseSensitiveColumnName(column.getFamilyName().getString(), column.getName().getString()), +column.getDataType().getSqlTypeName(), +column.isNullable(), +column.getMaxLength(), +column.getScale(), +false, +column.getSortOrder(), +column.getExpressionStr(), +column.isRowTimestamp()); +Expression defaultExpression = columnDef.getDefaultExpressionNode().accept(compiler); +if (!columnDef.getDefaultExpressionNode().isStateless() || !ExpressionUtil.isConstant(defaultExpression)) { +throw new SQLExceptionInfo.Builder(SQLExceptionCode.CANNOT_CREATE_STATEFUL_DEFAULT) + .setColumnName(columnDef.getColumnDefName().getColumnName()).build().buildException(); +} +column.setDefaultExpression(defaultExpression); +} +if (column.getDefaultExpression() != null) { +expression = new DefaultValueExpression(Arrays.asList(expression, column.getDefaultExpression())); +} +} {code} - Also, ColumnDef is immutable and needs to stay that way. Don't bother caching the Expression on it here: {code} +if (columnDef.getDefaultExpressionNode() != null) { +ExpressionCompiler compiler = new ExpressionCompiler(context); +Expression defaultExpression = columnDef.getDefaultExpressionNode().accept(compiler); +if (!columnDef.getDefaultExpressionNode().isStateless() || !ExpressionUtil.isConstant(defaultExpression)) { +throw new SQLExceptionInfo.Builder(SQLExceptionCode.CANNOT_CREATE_STATEFUL_DEFAULT) + .setColumnName(columnDef.getColumnDefName().getColumnName()).build().buildException(); +} +columnDef.setDefaultExpression(defaultExpression); +} {code} - Should be able to revert changes to PColumn and PColumnImpl (and sorry to have led you down that path). - There's a bit of code missing in PTableImpl.newKey() here (if this isn't the case, let me know what you mean by the "trailing null issue"): {code} public int newKey(ImmutableBytesWritable key, byte[][] values) { ... // If some non null pk values aren't set, then throw if (i < nColumns) { PColumn column = columns.get(i); if (column.getDataType().isFixedWidth() || !column.isNullable()) { throw new ConstraintViolationException(name.getString() + "." + column.getName().getString() + " may not be null"); } } {code} You'll need to loop until {{i == nColumns}} to check if the column is a PK column with a default value and then evaluate and set the value. You'll need to insert null values for PK columns which don't have values before that. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix >
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15556975#comment-15556975 ] James Taylor commented on PHOENIX-476: -- Thanks for the update, but I really need to see an updated patch at this point, [~kliew] to further advise you. The persistence is handled already - all you have to do is set the ColumnDef.expressionStr member variable. You can get the string version of a ParseNode using the {{parseNode.toString()}} method. This is what you need to set the expressionStr member variable to. Please make sure your patch covers these feedback items I already gave you: - Instead of using ExpressionUtil.isConstant(), use the following check: {code} expression.isStateless() && expression.getDeterminism() == Determinism.ALWAYS {code} - Don't wrap RowKeyExpression with getDefaultExpression for columns in the row key as we need to store the row key with the default value. To do this, change this code you added in ExpressionCompiler: {code} +if (column.getDefaultExpression() != null && !SchemaUtil.isPKColumn(column)) { +expression = new DefaultValueExpression(Arrays.asList(expression, column.getDefaultExpression())); +} {code} - Instead of ptr.copyBytes(), use ByteUtil.copyKeyBytesIfNecessary(ptr) as that prevents a copy when it's not necessary. - See https://phoenix.apache.org/building_website.html for how to update the website > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15556926#comment-15556926 ] Kevin Liew commented on PHOENIX-476: Hi [~jamestaylor], sorry for the delay. I've been working on this piecemeal. I've made changes to parse an expression string into a ParseNode so that default values will persist after a restart of the metadata server. I wanted to parse it in ColumnDef so we don't need to add another field to the constructor, but apparently expression string also has other uses for which it won't parse and will throw SQLException. I'll add the parsing to the ColumnDef constructor and catch the exception with the assumption that if an exception was thrown, the expression string was not intended as a default value string. Does that seem like a good approach? Regarding the case when nulls overwrite trailing default columns: the default value expression was not serialized as part of the projection, which explains why the coprocessors were returning null instead of the default value. I just need to wrap the expression with DefaultValueExpression in `ColumnRef.newColumnExpression` and do some verification. Can you point me to the location of the website/docs markdown pages? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=1482#comment-1482 ] James Taylor commented on PHOENIX-476: -- Any progress on this, [~kliew]? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15527244#comment-15527244 ] Kevin Liew commented on PHOENIX-476: Thanks James, will do. I didn’t have the right context; I was too focused on the individual parts but I’m starting to see how it fits together. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15524970#comment-15524970 ] James Taylor commented on PHOENIX-476: -- Please make the changes I recommended and let me know if there are any further issues (don't worry about any changes to {{PRowImpl.setValue}}). I've mentioned a couple time already that the {{KeyValueSchema.toBytes()}} code does not need to change because it executes *after* the default value would have been resolved on the server-side. We're only storing null values when the column is in the row key. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15524834#comment-15524834 ] Kevin Liew commented on PHOENIX-476: I did take a look at `PRowImpl.setValue` but that only determines how rows are serialized to disk. `KeyValueSchema.toBytes` determines how rows are serialized from the coprocessor to be sent to the driver. {code:java} if (expressions[index].evaluate(tuple, ptr) && ptr.getLength() > 0) { // Skip null values if (index >= minNullableIndex) { valueSet.set(index - minNullableIndex); } if (!type.isFixedWidth()) { b = ensureSize(b, offset, offset + getVarLengthBytes(ptr.getLength())); offset = writeVarLengthField(ptr, b, offset); } else { int nBytes = ptr.getLength(); b = ensureSize(b, offset, offset + nBytes); System.arraycopy(ptr.get(), ptr.getOffset(), b, offset, nBytes); offset += nBytes; } } {code} In the code above (for `KeyValueSchema.toBytes`), regardless of whether we store the null value on disk, nulls are not set in the value bitmask. The driver receives only non-null values and a value bitmask from the coprocessor. We could decide to store nulls but not default values on disk, which is what I think you were suggesting. Then we could differentiate nulls from default values and set default values in the value bitmask so that the driver will not see them as null values. But the driver also uses the value bitmask to determine the offset used to deserialize each cell, and now the driver will not be able to differentiate between non-null values that were serialized (and produce additional offset) and default values that were not serialized (and throws an exception because the expected byte array size is smaller than expected). The previous test cases were only passing before due to a quirk in the `KeyValueSchema.next` -> `Expression.evaluate` logic where non-trailing nulls evaluate false -> true while trailing nulls evaluate null -> false. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15523344#comment-15523344 ] James Taylor commented on PHOENIX-476: -- bq. I think that given the constraints that we have now, we might need to serialize another bitmask indicating default values. Does that sound like a good solution? There's no need to make any changes to this because the DefaultValueExpression would have already been executed and correctly determined the correct value (whether null or the default value). The ProjectedColumnExpression is evaluated on the client side, *after* the DefaultValueExpression(KeyValueExpression) would already have been evaluated, so if it's null at that point, it really should stay null. I think the remaining work is here[1]. Have you tried this yet, [~kliew] to see if there are any more issues once these are done? [1] https://issues.apache.org/jira/browse/PHOENIX-476?focusedCommentId=15507199=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15507199 > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15522376#comment-15522376 ] Kevin Liew commented on PHOENIX-476: Though I can see how this JIRA could be a large optimization for many use cases and it would be good to get this implemented sooner. Perhaps we should go with your original implementation idea while requiring that columns with a default value also have the `NOT NULL` constraint and implement additional DDL, before evaluating ways to provide support for nullable columns with default values. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=1551#comment-1551 ] Kevin Liew commented on PHOENIX-476: Thanks for the feedback guys. I investigated some more and I think I understand the issue more clearly now. We’ve set (in MetadataClient) tables with default value columns to store nulls in HBase and we can detect whether a cell is null or default in `PRowImpl.setValue`. Next, we’d need to communicate that from the coprocessor to the driver but the rows serialized on the coprocessors contain only the non-null values followed by a bitmask indicating set/null values. Currently both nulls and default values are indicated as nulls in the bitmask. We can’t change the value bitmask to indicate that default values are set and not null, because that would affect the size of the expected bytes and the offsets used for deserialization at the driver. I think that given the constraints that we have now, we might need to serialize another bitmask indicating default values. Does that sound like a good solution? I think we’d run into the same constraints after migrating to Calcite because the root of the issue lies in the communication between the coprocessor and the driver. Regarding `isConstant`: I think we can borrow the term ‘pure’ from functional programmers :). The current implementation for `isConstant` describes the criteria for a pure expression. Thanks for pointing out the parens [~julianhyde]. I’ll removed them so that the tests will pass after migrating to Calcite. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15507262#comment-15507262 ] James Taylor commented on PHOENIX-476: -- Yep, we'll disallow {{CURRENT_DATE}}. If we support it down the road, we'd have to persist the value for the reason Julian mentioned. bq. By the way, it's CURRENT_DATE not CURRENT_DATE(). Yeah, we're not following the standard - we should change it when we move to Calcite. SQL is weird, though - what's wrong with the parens... > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15507229#comment-15507229 ] Julian Hyde commented on PHOENIX-476: - I don't think you should allow {{CURRENT_DATE}} or similar. The standard says you should get the current date as of insert, but with your proposed implementation mechanism you will get the current date as of when the statement is executed. By the way, it's {{CURRENT_DATE}} not {{CURRENT_DATE()}}. SQL doesn't use parentheses for functions with zero arguments. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15507199#comment-15507199 ] James Taylor commented on PHOENIX-476: -- Thanks for the patch, [~kliew]. Nice work so far - it's coming along nicely. Here's some comments: - Instead of using ExpressionUtil.isConstant(), use the following check: {code} expression.isStateless() && expression.getDeterminism() == Determinism.ALWAYS {code} Maybe we should rename ExpressionUtil.isConstant() to ExpressionUtil.isConstantPerStatement() and use the above for a new ExpressionUtil.isConstant()? - We can't keep trailing nulls - that'll cause a world of problems. But on the plus side, since we put default values explicitly into rows, we don't need to wrap RowKeyExpression with getDefaultExpression. So you can just change this in ExpressionCompiler: {code} +if (column.getDefaultExpression() != null && !SchemaUtil.isPKColumn(column)) { +expression = new DefaultValueExpression(Arrays.asList(expression, column.getDefaultExpression())); +} {code} - Minor nit. Instead of ptr.copyBytes(), use ByteUtil.copyKeyBytesIfNecessary(ptr) as that prevents a copy when it's not necessary. - Not sure on the test failure, but I suspect you'll want to tweak the PRowImpl.setValue() call in PTableImpl. You'll want to be careful about the {{isNull}} setting and the case when there's a default value. You likely need to differentiate here between byteValue of null and a byteValue of an empty byteValue: {code} public void setValue(PColumn column, byte[] byteValue) { deleteRow = null; byte[] family = column.getFamilyName().getBytes(); byte[] qualifier = column.getName().getBytes(); PDataType type = column.getDataType(); // Check null, since some types have no byte representation for null boolean isNull = type.isNull(byteValue); // TODO: this might need to change {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.3.patch, > PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15504528#comment-15504528 ] James Taylor commented on PHOENIX-476: -- Yep, Phoenix can distinguish between these two cases, so should be fine. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15504513#comment-15504513 ] Julian Hyde commented on PHOENIX-476: - [~jamestaylor], I'm skeptical. DEFAULT is clearly intended to be implemented when data is written, not when it is read. If HBase is able to distinguish missing from null values then it's possible that your plan would work, and you can make Phoenix consistent with standard semantics without storing millions of values that are the same as the default value. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15501243#comment-15501243 ] James Taylor commented on PHOENIX-476: -- Thanks for chiming in, [~julianhyde]. So we can still implement DEFAULT without storing the default value for non PK columns, [~kliew]. This is important as for a multi billion row table, it will save a lot of space. Instead of using using CoalesceFunction during the wrapping, we can create a new Expression and register it in ExpressionType, called DefaultValueExpression. The evaluate method can distinguish between the case where there's no value for the column (i.e. colRefChildExpression.evaluate returns false) versus it returning a null value (i.e. colRefChildExpression.evaluate returns true but ptr.getLength() is zero). We'd want to force STORE_NULLS=true when a table is created with default values. See MetaDataClient.createTableInternal() and how we force this option to true for transactional tables - we'd want to do something similar here if any of the columns for the new table define a DEFAULT. With this option in place, Phoenix will store an empty byte array as the column value instead of issuing an HBase delete. The evaluate method of DefaultValueExpression would look something like this (just one very subtle change from the CoalesceFunction implementation): {code} @Override public boolean evaluate(Tuple tuple, ImmutableBytesWritable ptr) { boolean evaluated = children.get(0).evaluate(tuple, ptr); if (evaluated) { // Will potentially evaluate to null without evaluating the second expression return true; } if (tuple.isImmutable()) { // This is true if it's the last time an evaluation is happening on the row Expression secondChild = children.get(1); if (secondChild.evaluate(tuple, ptr)) { // Coerce the type of the second child to the type of the first child getDataType().coerceBytes(ptr, secondChild.getDataType(), secondChild.getSortOrder(), getSortOrder()); return true; } } return false; } {code} > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15497639#comment-15497639 ] Julian Hyde commented on PHOENIX-476: - I agree with [~kliew]. The result should be (1, NULL). {{DEFAULT}} kicks in only in INSERT or UPSERT, and only when the column is not mentioned. If you insert a NULL value, the value becomes NULL. The other way to get a default value is to mention the column explicitly and use DEFAULT, e.g. {{INSERT INTO testdefault (pk, test) VALUES (1, DEFAULT)}}. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15497545#comment-15497545 ] Kevin Liew commented on PHOENIX-476: The SQL92 spec states: {noformat} 4) Let Q be the result of that . Case: a) If Q is empty, then no row is inserted and a completion con- dition is raised: no data. b) Otherwise, for each row R of Q: i) A candidate row of B is effectively created in which the value of each column is its default value, as specified in the General Rules of Subclause 11.5, "". The candidate row includes every column of B. ii) For every object column in the candidate row, the value of the object column identified by the i-th in the is replaced by the i-th value of R. iii) Let C be a column that is represented in the candidate row and let SV be its value in the candidate row. The General Rules of Subclause 9.2, "Store assignment", are applied to C and SV as TARGET and VALUE, respectively. iv) The candidate row is inserted into B. Note: The data values allowable in the candidate row may be constrained by a WITH CHECK OPTION constraint. The effect of a WITH CHECK OPTION constraint is defined in the General Rules of Subclause 11.19, "". {noformat} In step `b.i`, a candidate row with all default values will be constructed. Step `b.ii` is a bit ambiguous but I'd interpret that step such that a default value value could be overwritten with a null value. I tested with MySQL and Postgres and they both allow overwriting a default with a null value. Any input [~julianhyde]? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15497244#comment-15497244 ] James Taylor commented on PHOENIX-476: -- If you set a column value with a default back to null, then you should get the default value when you query it. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15497128#comment-15497128 ] Kevin Liew commented on PHOENIX-476: If we consider the following case: {noformat} create table testdefault (pk integer primary key, test integer default 5); upsert into testdefault values (1, null); select * from testdefault; {noformat} We expect `1, ` but if we use `COALESCE` while retrieving the second column, would we get `1, 5`? > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15497020#comment-15497020 ] James Taylor commented on PHOENIX-476: -- Thanks for the updated patch, [~kliew]. bq. For non-pk columns, if we evaluate the expression once on and wrap the column in `COALESCE` to replace `NULL` values with the expression See my bullet point above starting with "For other, non PK columns, you don't need to store the value at all". You wouldn't evaluate the expression. You're replacing an expression of {{col1}} with {{COALESCE(col1,)}}. This will do the right thing. No need to write the default value for non pk columns. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.2.patch, PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15475294#comment-15475294 ] James Taylor commented on PHOENIX-476: -- Thanks for the patch, [~kliew]. FYI, a fair amount of the work is already done thanks to functional indexes. Here's some feedback: - In the grammar file, allow DEFAULT to be any expression (i.e. it'll end up being a ParseNode). - Set the ColumnDef expressionStr argument based on the defaultNode.toString(). We could potentially add a new ColumnDef constructor that takes a ParseNode instead of String and a getExpressionNode() getter. You could check here that defaultNode.isStateless() is true and raise an exception if it isn't. - In CreateTableCompiler, compile ColumnDef.getExpressionNode() using an ExpressionCompiler and verify that defaultExpression.isStateless() and defaultExpression.getDeterminism() == Determinism.ALWAYS. The latter will rule out CURRENT_DATE() and NEXT VALUE FOR expressions. - For PColumn, we already have getExpressionStr() which is all we need for now. We might want to add the following to cache the Expression version of the same {code} Expression getDefaultExpression(PhoenixConnection conn); {code} The getDefaultExpression will be similar to the PTable.getIndexMaintainer call in that it'll parse and compile the string and cache it on PColumnImpl. - The default value will be persisted in SYSTEM.CATALOG in the COLUMN_DEF column already based on ColumnDef.getExpressionStr(). It'll also flow already between client and server in PColumn (which is great). - Instead of always setting the default value in UpsertCompiler for UPSERT VALUE, you can only set it for row key columns. This logic can live in PTable.newKey() method - just use the default value if a column value is null which has a default. You can use ExpressionUtil.getConstantExpression() to evaluate it dynamically to get the actual bytes to use (which will be returned in ptr). - For other, non PK columns, you don't need to store the value at all (see discussion above). Instead, in ExpressionCompiler.resolveColumn(), before returning {{ref}} (but only if it's not a PK column), you'd check if the PColumn has a default value using PColumn.getExpressionStr(). If it does, you'd wrap the {{ref}} in a CoalesceFunction with the first child being {{ref}} and the second child being the defaultExpression. In this way, when evaluated, we'd fall back to the default value if the column has no value. - For indexes, copy over the expressionStr from the PColumn of the data table in MetaDataClient.createIndex() when we're creating the ColumnDef for covered columns (might already be occurring). One open question, using this technique, what if the default value changes? Any thoughts on this, [~lhofhansl]? Since we didn't persist it before, rows that already exist would get the new default value. We don't have a mechanism for it to be changed now, so it's a bit of a theoretical issue currently. I suppose in theory we could persist the old value if the value were to change. Or lookup the default value for the column based on the timestamp of the data cell (which would get complicated). > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor >Assignee: Kevin Liew > Labels: enhancement > Attachments: PHOENIX-476.patch > > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15459018#comment-15459018 ] James Taylor commented on PHOENIX-476: -- [~kliew] - would you be interested in this one? We could do it in two stages: #. Allow only constants (i.e. ParseNode.isStateless() of true). #. Allow things like CURRENT_DATE and NEXT VALUE FOR expressions which evaluate to constants, but must be evaluated per statement so they're a bit trickier. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor > Labels: enhancement > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15367829#comment-15367829 ] James Heather commented on PHOENIX-476: --- Is there any progress on this? It would be really useful to have default values. > Support declaration of DEFAULT in CREATE statement > -- > > Key: PHOENIX-476 > URL: https://issues.apache.org/jira/browse/PHOENIX-476 > Project: Phoenix > Issue Type: Task >Affects Versions: 3.0-Release >Reporter: James Taylor > Labels: enhancement > > Support the declaration of a default value in the CREATE TABLE/VIEW statement > like this: > CREATE TABLE Persons ( > Pid int NOT NULL PRIMARY KEY, > LastName varchar(255) NOT NULL, > FirstName varchar(255), > Address varchar(255), > City varchar(255) DEFAULT 'Sandnes' > ) > To implement this, we'd need to: > 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through > the value when the table is created (in MetaDataClient). > 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value > is present, since the column will never be null. > 3. add a getDefaultValue() accessor in PColumn > 4. for a row key column, during UPSERT use the default value if no value was > specified for that column. This could be done in the PTableImpl.newKey method. > 5. for a key value column with a default value, we can get away without > incurring any storage cost. Although a little bit of extra effort than if we > persisted the default value on an UPSERT for key value columns, this approach > has the benefit of not incurring any storage cost for a default value. > * serialize any default value into KeyValueColumnExpression > * in the evaluate method of KeyValueColumnExpression, conditionally use > the default value if the column value is not present. If doing partial > evaluation, you should not yet return the default value, as we may not have > encountered the the KeyValue for the column yet (since a filter evaluates > each time it sees each KeyValue, and there may be more than one KeyValue > referenced in the expression). Partial evaluation is determined by calling > Tuple.isImmutable(), where false means it is NOT doing partial evaluation, > while true means it is. > * modify EvaluateOnCompletionVisitor by adding a visitor method for > RowKeyColumnExpression and KeyValueColumnExpression to set > evaluateOnCompletion to true if they have a default value specified. This > will cause filter evaluation to execute one final time after all KeyValues > for a row have been seen, since it's at this time we know we should use the > default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14117125#comment-14117125 ] Gabriel Reid commented on PHOENIX-476: -- [~jamestaylor] yes, definitely looks useful for bulk loading scenarios. I'll try to take a look at this one in the next little big, but I won't be able to get to it right away. It looks like the hard part (i.e. figuring out how to actually do it) is pretty much done, so hopefully the impl work shouldn't be too bad. Support declaration of DEFAULT in CREATE statement -- Key: PHOENIX-476 URL: https://issues.apache.org/jira/browse/PHOENIX-476 Project: Phoenix Issue Type: Task Affects Versions: 3.0-Release Reporter: James Taylor Labels: enhancement Support the declaration of a default value in the CREATE TABLE/VIEW statement like this: CREATE TABLE Persons ( Pid int NOT NULL PRIMARY KEY, LastName varchar(255) NOT NULL, FirstName varchar(255), Address varchar(255), City varchar(255) DEFAULT 'Sandnes' ) To implement this, we'd need to: 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through the value when the table is created (in MetaDataClient). 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value is present, since the column will never be null. 3. add a getDefaultValue() accessor in PColumn 4. for a row key column, during UPSERT use the default value if no value was specified for that column. This could be done in the PTableImpl.newKey method. 5. for a key value column with a default value, we can get away without incurring any storage cost. Although a little bit of extra effort than if we persisted the default value on an UPSERT for key value columns, this approach has the benefit of not incurring any storage cost for a default value. * serialize any default value into KeyValueColumnExpression * in the evaluate method of KeyValueColumnExpression, conditionally use the default value if the column value is not present. If doing partial evaluation, you should not yet return the default value, as we may not have encountered the the KeyValue for the column yet (since a filter evaluates each time it sees each KeyValue, and there may be more than one KeyValue referenced in the expression). Partial evaluation is determined by calling Tuple.isImmutable(), where false means it is NOT doing partial evaluation, while true means it is. * modify EvaluateOnCompletionVisitor by adding a visitor method for RowKeyColumnExpression and KeyValueColumnExpression to set evaluateOnCompletion to true if they have a default value specified. This will cause filter evaluation to execute one final time after all KeyValues for a row have been seen, since it's at this time we know we should use the default value. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (PHOENIX-476) Support declaration of DEFAULT in CREATE statement
[ https://issues.apache.org/jira/browse/PHOENIX-476?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=14116174#comment-14116174 ] James Taylor commented on PHOENIX-476: -- [~gabriel.reid] - interested in this one? It would be very useful to be able to declare an auto increment field in the DDL as follows to support use case like this one: http://s.apache.org/Ur1 {code} CREATE TABLE my_table ( my_pk BIGINT PRIMARY_KEY DEFAULT NEXT VALUE FOR my_seq, my_col VARCHAR ) {code} Support declaration of DEFAULT in CREATE statement -- Key: PHOENIX-476 URL: https://issues.apache.org/jira/browse/PHOENIX-476 Project: Phoenix Issue Type: Task Affects Versions: 3.0-Release Reporter: James Taylor Labels: enhancement Support the declaration of a default value in the CREATE TABLE/VIEW statement like this: CREATE TABLE Persons ( Pid int NOT NULL PRIMARY KEY, LastName varchar(255) NOT NULL, FirstName varchar(255), Address varchar(255), City varchar(255) DEFAULT 'Sandnes' ) To implement this, we'd need to: 1. add a new DEFAULT_VALUE key value column in SYSTEM.TABLE and pass through the value when the table is created (in MetaDataClient). 2. always set NULLABLE to ResultSetMetaData.columnNoNulls if a default value is present, since the column will never be null. 3. add a getDefaultValue() accessor in PColumn 4. for a row key column, during UPSERT use the default value if no value was specified for that column. This could be done in the PTableImpl.newKey method. 5. for a key value column with a default value, we can get away without incurring any storage cost. Although a little bit of extra effort than if we persisted the default value on an UPSERT for key value columns, this approach has the benefit of not incurring any storage cost for a default value. * serialize any default value into KeyValueColumnExpression * in the evaluate method of KeyValueColumnExpression, conditionally use the default value if the column value is not present. If doing partial evaluation, you should not yet return the default value, as we may not have encountered the the KeyValue for the column yet (since a filter evaluates each time it sees each KeyValue, and there may be more than one KeyValue referenced in the expression). Partial evaluation is determined by calling Tuple.isImmutable(), where false means it is NOT doing partial evaluation, while true means it is. * modify EvaluateOnCompletionVisitor by adding a visitor method for RowKeyColumnExpression and KeyValueColumnExpression to set evaluateOnCompletion to true if they have a default value specified. This will cause filter evaluation to execute one final time after all KeyValues for a row have been seen, since it's at this time we know we should use the default value. -- This message was sent by Atlassian JIRA (v6.2#6252)