chore: Add hive windows workflow#4850
Conversation
kazuyukitanimura
left a comment
kazuyukitanimura
left a comment
There was a problem hiding this comment.
Just one comment
pending ci
| # not include it, so `hive/testOnly *` above never sees HiveWindowFunctionQuerySuite | ||
| # et al. Prepend an SBT `set` command to graft that source directory into the hive | ||
| # project's test sources, then target the three compatibility suites explicitly. | ||
| - {name: "sql_hive_compat-1", args1: "'set LocalProject(\"hive\") / Test / unmanagedSourceDirectories += file(\"sql/hive/compatibility/src/test/scala\")'", args2: "hive/testOnly org.apache.spark.sql.hive.execution.HiveWindowFunctionQuerySuite org.apache.spark.sql.hive.execution.HiveWindowFunctionQueryFileSuite org.apache.spark.sql.hive.execution.HiveCompatibilitySuite -- -l org.apache.spark.tags.ExtendedHiveTest -l org.apache.spark.tags.SlowHiveTest"} |
There was a problem hiding this comment.
Should we use hive/test instead of hive/testOnly and exclude already tested tags...
|
Nice, I see it reproduced the issue |
|
The Spark checks Spark result against HiveQL, compares the string by string, no way we can insert something like the golden file looks like that So I'm not sure if this hive WF actually needed |
If we decide not to skip these tests then it should be possible to change Spark's comparison to be smarter in the diff file. |
4 participants
Which issue does this PR close?
Closes #.
Rationale for this change
What changes are included in this PR?
How are these changes tested?