Author: gates
Date: Sat Mar 3 22:13:09 2012
New Revision: 1296714
URL: http://svn.apache.org/viewvc?rev=1296714&view=rev
Log:
HCATALOG-264 Barrier for HCatOutputFormat for Pig/Mapreduce if the table
contains feature only supported in Hive
Modified:
incubator/hcatalog/branches/branch-0.4/CHANGES.txt
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/common/ErrorType.java
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/mapreduce/HCatOutputFormat.java
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/drivers/TestDriverPig.pm
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf
incubator/hcatalog/branches/branch-0.4/src/test/org/apache/hcatalog/mapreduce/TestHCatOutputFormat.java
Modified: incubator/hcatalog/branches/branch-0.4/CHANGES.txt
URL:
http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/CHANGES.txt?rev=1296714&r1=1296713&r2=1296714&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/CHANGES.txt (original)
+++ incubator/hcatalog/branches/branch-0.4/CHANGES.txt Sat Mar 3 22:13:09 2012
@@ -21,7 +21,9 @@ Apache HCatalog Change Log
Release 0.4.0 - Unreleased
INCOMPATIBLE CHANGES
- HCAT-279 Remove remaining code mentioning isd/osd (diajy via gates)
+ HCAT-264 Barrier for HCatOutputFormat for Pig/Mapreduce if the table
contains feature only supported in Hive (daijy via gates)
+
+ HCAT-279 Remove remaining code mentioning isd/osd (daijy via gates)
HCAT-261 Enable more DDL statements for HCat (daijy via gates)
Modified:
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/common/ErrorType.java
URL:
http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/common/ErrorType.java?rev=1296714&r1=1296713&r2=1296714&view=diff
==============================================================================
---
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/common/ErrorType.java
(original)
+++
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/common/ErrorType.java
Sat Mar 3 22:13:09 2012
@@ -45,6 +45,7 @@ public enum ErrorType {
ERROR_TOO_MANY_DYNAMIC_PTNS (2013, "Attempt to create too many
dynamic partitions"),
ERROR_INIT_LOADER (2014, "Error initializing Pig
loader"),
ERROR_INIT_STORER (2015, "Error initializing Pig
storer"),
+ ERROR_NOT_SUPPORTED (2016, "Error operation not
supported"),
/* Authorization Errors 3000 - 3999 */
ERROR_ACCESS_CONTROL (3000, "Permission denied"),
Modified:
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/mapreduce/HCatOutputFormat.java
URL:
http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/mapreduce/HCatOutputFormat.java?rev=1296714&r1=1296713&r2=1296714&view=diff
==============================================================================
---
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/mapreduce/HCatOutputFormat.java
(original)
+++
incubator/hcatalog/branches/branch-0.4/src/java/org/apache/hcatalog/mapreduce/HCatOutputFormat.java
Sat Mar 3 22:13:09 2012
@@ -34,6 +34,7 @@ import org.apache.hadoop.hive.conf.HiveC
import org.apache.hadoop.hive.conf.HiveConf.ConfVars;
import org.apache.hadoop.hive.metastore.HiveMetaStoreClient;
import org.apache.hadoop.hive.metastore.api.FieldSchema;
+import org.apache.hadoop.hive.metastore.api.Index;
import org.apache.hadoop.hive.metastore.api.MetaException;
import org.apache.hadoop.hive.metastore.api.StorageDescriptor;
import org.apache.hadoop.hive.metastore.api.Table;
@@ -77,7 +78,29 @@ public class HCatOutputFormat extends HC
hiveConf = HCatUtil.getHiveConf(null, conf);
client = HCatUtil.createHiveClient(hiveConf);
Table table = client.getTable(outputJobInfo.getDatabaseName(),
outputJobInfo.getTableName());
-
+
+ List<String> indexList =
client.listIndexNames(outputJobInfo.getDatabaseName(),
outputJobInfo.getTableName(), Short.MAX_VALUE);
+
+ for (String indexName : indexList) {
+ Index index = client.getIndex(outputJobInfo.getDatabaseName(),
outputJobInfo.getTableName(), indexName);
+ if (!index.isDeferredRebuild()) {
+ throw new HCatException(ErrorType.ERROR_NOT_SUPPORTED, "Store
into a table with an automatic index from Pig/Mapreduce is not supported");
+ }
+ }
+ StorageDescriptor sd = table.getSd();
+
+ if (sd.isCompressed()) {
+ throw new HCatException(ErrorType.ERROR_NOT_SUPPORTED, "Store into
a compressed partition from Pig/Mapreduce is not supported");
+ }
+
+ if (sd.getBucketCols()!=null && !sd.getBucketCols().isEmpty()) {
+ throw new HCatException(ErrorType.ERROR_NOT_SUPPORTED, "Store into
a partition with bucket definition from Pig/Mapreduce is not supported");
+ }
+
+ if (sd.getSortCols()!=null && !sd.getSortCols().isEmpty()) {
+ throw new HCatException(ErrorType.ERROR_NOT_SUPPORTED, "Store into
a partition with sorted column definition from Pig/Mapreduce is not supported");
+ }
+
if (table.getPartitionKeysSize() == 0 ){
if ((outputJobInfo.getPartitionValues() != null) &&
(!outputJobInfo.getPartitionValues().isEmpty())){
// attempt made to save partition values in non-partitioned table
- throw error.
Modified:
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/drivers/TestDriverPig.pm
URL:
http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/drivers/TestDriverPig.pm?rev=1296714&r1=1296713&r2=1296714&view=diff
==============================================================================
---
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/drivers/TestDriverPig.pm
(original)
+++
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/drivers/TestDriverPig.pm
Sat Mar 3 22:13:09 2012
@@ -192,7 +192,7 @@ sub runTest
# run a another Pig script to dump the results of the table.
my $result;
if (defined($testCmd->{'result_table'})) {
- $result = $self->runPig( $testCmd, $log, 0);
+ $result = $self->runPig( $testCmd, $log, 0, 1);
my @results = ();
my @outputs = ();
if (ref($testCmd->{'result_table'}) ne 'ARRAY') {
@@ -214,7 +214,7 @@ sub runTest
$tableName = $results[$i];
$modifiedTestCmd{'num'} = $testCmd->{'num'} . "_" . $i .
"_benchmark";
$modifiedTestCmd{'pig'} = "a = load '$tableName' using
org.apache.hcatalog.pig.HCatLoader(); store a into ':OUTPATH:';";
- my $r = $self->runPig(\%modifiedTestCmd, $log, 1);
+ my $r = $self->runPig(\%modifiedTestCmd, $log, 1, 1);
$outputs[$i] = $r->{'output'};
} else {
$localdir = $testCmd->{'localpath'} . $testCmd->{'group'} .
"_" . $testCmd->{'num'} . ".out/$id";
@@ -233,7 +233,7 @@ sub runTest
}
}
else {
- $result = $self->runPig( $testCmd, $log, 1);
+ $result = $self->runPig( $testCmd, $log, 1, 1);
}
return $result;
} elsif( $testCmd->{'script'} ){
@@ -447,7 +447,7 @@ sub dumpPigTable
sub runPig
{
- my ($self, $testCmd, $log, $copyResults) = @_;
+ my ($self, $testCmd, $log, $copyResults, $noFailOnFail) = @_;
my $subName = (caller(0))[3];
my %result;
@@ -488,8 +488,13 @@ sub runPig
print $log "Setting PIG_CLASSPATH to $ENV{'PIG_CLASSPATH'}\n";
print $log "$0::$className::$subName INFO: Going to run pig command:
@cmd\n";
- IPC::Run::run(\@cmd, \undef, $log, $log) or
+ my $runrc = IPC::Run::run(\@cmd, \undef, $log, $log);
+
+ if (defined($noFailOnFail) && $noFailOnFail) {
+ } else {
die "Failed running $pigfile\n";
+ }
+
$result{'rc'} = $? >> 8;
Modified:
incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf
URL:
http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf?rev=1296714&r1=1296713&r2=1296714&view=diff
==============================================================================
--- incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf
(original)
+++ incubator/hcatalog/branches/branch-0.4/src/test/e2e/hcatalog/tests/pig.conf
Sat Mar 3 22:13:09 2012
@@ -280,6 +280,28 @@ store a into ':OUTPATH:';\,
}
],
}, # end g
-
+ {
+ 'name' => 'Pig_HCat_Barrier',
+ 'tests' => [
+ {
+ 'num' => 1
+ ,'hcat_prep'=>q\drop table if exists
pig_hcat_barrier_1;
+create table pig_hcat_barrier_1 (name string, age int, gpa double) partitioned
by (b string) CLUSTERED BY (name) INTO 1 BUCKETS STORED AS TEXTFILE;\
+ ,'pig' => q\
+a = load ':INPATH:/studenttab10k' as (name:chararray, age:int, gpa:double);
+store a into 'pig_hcat_barrier_1' using
org.apache.hcatalog.pig.HCatStorer('b=1'); \,
+ ,'expected_err_regex' => 'not supported'
+ },
+ {
+ 'num' => 2
+ ,'hcat_prep'=>q\drop table if exists
pig_hcat_barrier_2;
+create table pig_hcat_barrier_2 (name string, age int, gpa double) partitioned
by (b string) CLUSTERED BY (name) SORTED BY (name) INTO 1 BUCKETS STORED AS
TEXTFILE;\
+ ,'pig' => q\
+a = load ':INPATH:/studenttab10k' as (name:chararray, age:int, gpa:double);
+store a into 'pig_hcat_barrier_2' using
org.apache.hcatalog.pig.HCatStorer('b=1'); \,
+ ,'expected_err_regex' => 'not supported'
+ },
+ ],
+ }, # end g
]
}
Modified:
incubator/hcatalog/branches/branch-0.4/src/test/org/apache/hcatalog/mapreduce/TestHCatOutputFormat.java
URL:
http://svn.apache.org/viewvc/incubator/hcatalog/branches/branch-0.4/src/test/org/apache/hcatalog/mapreduce/TestHCatOutputFormat.java?rev=1296714&r1=1296713&r2=1296714&view=diff
==============================================================================
---
incubator/hcatalog/branches/branch-0.4/src/test/org/apache/hcatalog/mapreduce/TestHCatOutputFormat.java
(original)
+++
incubator/hcatalog/branches/branch-0.4/src/test/org/apache/hcatalog/mapreduce/TestHCatOutputFormat.java
Sat Mar 3 22:13:09 2012
@@ -109,8 +109,8 @@ public class TestHCatOutputFormat extend
sd.setOutputFormat(RCFileOutputFormat.class.getName());
sd.setParameters(new HashMap<String, String>());
sd.getParameters().put("test_param_1", "Use this for comments etc");
- sd.setBucketCols(new ArrayList<String>(2));
- sd.getBucketCols().add("name");
+ //sd.setBucketCols(new ArrayList<String>(2));
+ //sd.getBucketCols().add("name");
sd.setSerdeInfo(new SerDeInfo());
sd.getSerdeInfo().setName(tbl.getTableName());
sd.getSerdeInfo().setParameters(new HashMap<String, String>());
