020b2622e5
### What changes were proposed in this pull request? After https://github.com/apache/spark/pull/27387 (see https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-hive-2.3/202/), the tests below fail consistently, specifically in one job https://amplab.cs.berkeley.edu/jenkins/job/spark-master-test-sbt-hadoop-2.7-hive-2.3/ in Jenkins ``` org.apache.spark.sql.hive.HiveShowCreateTableSuite.simple hive table org.apache.spark.sql.hive.HiveShowCreateTableSuite.simple external hive table org.apache.spark.sql.hive.HiveShowCreateTableSuite.hive bucketing is supported ``` The profile is same as PR builder but seems it fails specifically in this machine. Seems the legacy configuration `spark.sql.legacy.createHiveTableByDefault.enabled` is not being set due to the inconsistency between `SQLConf.get` and the active Spark session as described in the https://github.com/apache/spark/pull/27387. This PR proposes to explicitly set the configuration against the session used instead of `SQLConf.get`. ### Why are the changes needed? To make `spark-master-test-sbt-hadoop-2.7-hive-2.3` job pass. ### Does this PR introduce any user-facing change? No. ### How was this patch tested? Cannot reproduce in my local. Presumably it cannot be reproduced in the PR builder. We should see if the tests pass at `spark-master-test-sbt-hadoop-2.7-hive-2.3` job after this PR is merged Closes #27703 from HyukjinKwon/SPARK-30798-followup. Authored-by: HyukjinKwon <gurwls223@apache.org> Signed-off-by: HyukjinKwon <gurwls223@apache.org> |
||
---|---|---|
.. | ||
benchmarks | ||
compatibility/src/test/scala/org/apache/spark/sql/hive/execution | ||
src | ||
pom.xml |