[
https://issues.apache.org/jira/browse/PHOENIX-5707?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17027924#comment-17027924
]
Swaroopa Kadam commented on PHOENIX-5707:
-----------------------------------------
[~priyankporwal] Thank you for the help in reproducing the problem!
> Index rebuild after truncate incorrectly writes the included column value
> -------------------------------------------------------------------------
>
> Key: PHOENIX-5707
> URL: https://issues.apache.org/jira/browse/PHOENIX-5707
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.15.0
> Reporter: Swaroopa Kadam
> Assignee: Swaroopa Kadam
> Priority: Minor
> Fix For: 5.1.0, 4.15.1
>
>
>
> {code:java}
> @Test
> public void testIncorrectRebuild() throws Exception {
> String schemaName = generateUniqueName();
> String dataTableName = generateUniqueName();
> String dataTableFullName = SchemaUtil.getTableName(schemaName,
> dataTableName);
> String indexTableName = generateUniqueName();
> String indexTableFullName = SchemaUtil.getTableName(schemaName,
> indexTableName);
> Properties props = PropertiesUtil.deepCopy(TEST_PROPERTIES);
> try (Connection conn = DriverManager.getConnection(getUrl(), props)) {
> conn.setAutoCommit(true);
> conn.createStatement().execute("create table " + dataTableFullName +
> " (id varchar(10) not null primary key, val1 varchar(10),
> val2 varchar(10), val3 varchar(10)) COLUMN_ENCODED_BYTES=0");
> conn.createStatement().execute("CREATE INDEX " + indexTableName + "
> on " +
> dataTableFullName + " (val1) include (val2, val3)" );
> //insert a full row
> conn.createStatement().execute("upsert into " + dataTableFullName + "
> values ('a', 'ab', 'efgh', 'abcd')");
> Thread.sleep(1000);
> // insert a partial row
> conn.createStatement().execute("upsert into " + dataTableFullName + "
> (id, val3) values ('a', 'uvwx')");
> Thread.sleep(1000);
> //insert a full row
> conn.createStatement().execute("upsert into " + dataTableFullName + "
> values ('a', 'ab', 'efgh', 'yuio')");
> Thread.sleep(1000);
> //insert a partial row
> conn.createStatement().execute("upsert into " + dataTableFullName + "
> (id, val3) values ('a', 'asdf')");
> //truncate index table
> ConnectionQueryServices queryServices =
> conn.unwrap(PhoenixConnection.class).getQueryServices();
> Admin admin = queryServices.getAdmin();
> TableName tableName = TableName.valueOf(indexTableFullName);
> admin.disableTable(tableName);
> admin.truncateTable(tableName, false);
> //rebuild index
> runIndexTool(true, false, schemaName, dataTableName, indexTableName);
> // we expect 2 versions to be written after rebuild, one for last full
> row update and one for latest update
> //assert
> Scan scan = new Scan();
> scan.setRaw(true);
> scan.setMaxVersions(10);
> HTable indexTable = new HTable(getUtility().getConfiguration(),
> indexTableFullName);
> HTable dataTable = new HTable(getUtility().getConfiguration(),
> dataTableFullName);
> long dataFullRowTS = 0;
> ResultScanner rs = dataTable.getScanner(scan);
> for (Result r : rs) {
> for (Cell c : r.listCells()) {
> String column = new String(CellUtil.cloneQualifier(c));
> String value = new String(CellUtil.cloneValue(c));
> if (column.equalsIgnoreCase("VAL3") &&
> value.equalsIgnoreCase("yuio")) {
> dataFullRowTS = c.getTimestamp();
> }
> }
> }
> rs = indexTable.getScanner(scan);
> for (Result r : rs) {
> for (Cell c : r.listCells()) {
> long indexTS = c.getTimestamp();
> String column = new String(CellUtil.cloneQualifier(c));
> if (column.equalsIgnoreCase("0:VAL3") && indexTS ==
> dataFullRowTS) {
> String value = new String(CellUtil.cloneValue(c));
> assertEquals("yuio", value); // if the ts is from full
> rebuild row , value should also be from full rebuild row
> }
> }
> }
> }
> }
> {code}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
