Upgrade Hive JDBC version to 3.1.3 #5
Conversation
Otherwise, it throws an IllegalArgumentException.
The previous version 0.13.0 throws an exception when executing an query on Spark 3.4.0.
There was a problem hiding this comment.
Thanks for updating this. We need to either add more exclusions or adjust the shading rules. If you look at the current release JAR, everything is under one of these:
io/trino/hive/jdbc/$internal
org/apache/hive/jdbc
The version from this PR has lots of extra stuff, which you can see by running the following:
jar tf target/hive-apache-jdbc-*-SNAPSHOT.jar | sort
I suspect that we can exclude most of these things. It might help to look at the exclusion rules from our trino-hive-apache project, as those will match the current Hive version.
| <dep.guava.version>11.0.2</dep.guava.version> | ||
| <dep.slf4j.version>1.7.10</dep.slf4j.version> | ||
| <dep.hadoop.version>1.2.1</dep.hadoop.version> | ||
| <dep.hive.version>3.1.3</dep.hive.version> |
There was a problem hiding this comment.
Update the project version to 3.1.3-1-SNAPSHOT so that it matches the Hive version.
Update the README to use the new version, or simply remove the Hive version from the README.
|
Hmm, it seems the new Hive JDBC driver can't connect to the old Hive server (e.g. |
|
Installing standalone Hive JDBC driver at runtime like trinodb/trino#17667 looks better because we need to use two JDBC versions for connecting to old Hive server and new Spark server. |
The previous version 0.13.0 throws an exception
when executing a query on Spark 3.4.0.
Relates to trinodb/trino#17667