[
https://issues.apache.org/jira/browse/SPARK-59785?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-59785:
-----------------------------------
Labels: pull-request-available (was: )
> Fix SparkConnectResultSetSuite failure when ANSI mode is off
> ------------------------------------------------------------
>
> Key: SPARK-59785
> URL: https://issues.apache.org/jira/browse/SPARK-59785
> Project: Spark
> Issue Type: Test
> Components: Connect
> Affects Versions: 5.0.0
> Reporter: sepuri sai krishna
> Priority: Major
> Labels: pull-request-available
>
> {{SparkConnectResultSetSuite}} test "next throws SQLException when the query
> fails during result streaming", added in SPARK-57933, fails in every Non-ANSI
> scheduled build (master, branch-4.x and branch-4.3):
> {noformat}
> - next throws SQLException when the query fails during result streaming ***
> FAILED ***
> Expected exception java.sql.SQLException to be thrown, but no exception was
> thrown (SparkConnectResultSetSuite.scala:195)
> {noformat}
> The test runs {{SELECT 10 / (5 - id) FROM range(10)}} and expects a
> {{DIVIDE_BY_ZERO}} error (SQLSTATE 22012). That error is only raised when
> ANSI mode is on. The Non-ANSI builds set {{SPARK_ANSI_SQL_MODE=false}}, so
> the division returns NULL and nothing is thrown.
> Reproduce with:
> {noformat}
> SPARK_ANSI_SQL_MODE=false build/sbt 'connect-client-jdbc/testOnly
> *SparkConnectResultSetSuite'
> {noformat}
> Proposed fix: enable {{spark.sql.ansi.enabled}} explicitly for the test's
> JDBC session. Each JDBC connection uses its own Spark Connect session, so the
> setting does not leak to other tests.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
