zhouyifan279 commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-2100506618
@jeanlyn
> 2. Use partitionPaths to get matchingPartitions, then get
customPartitionLocations like what we do in this PR.
We can create a function `patitionPaths:
jeanlyn commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-2100456683
@zhouyifan279 , I think the solution you mention can solve the data
inconsistency for custom partitions. But i don't know it's acceptable to
communicate to metastore in such low
zhouyifan279 commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-2100365610
To eliminate data inconsistency issue, we should handle custom partitions in
`HadoopMapReduceCommitProtocol.commitJob` instead of writing to the final
output path then moving
bowenliang123 commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-1903692866
+1 for this PR. This PR is worth a second chance to be re-evaluated and it
is a reasonable approach to reduce the number of requests and request pressure
to Hive metastore service.
jeanlyn commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-1903655119
Hi, @npsables , we have using this patch running in our production for
several months, you can try this patch safely. I don't think the community will
merge this patch, because the
npsables commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-1903626788
Hi guys, will this ever get merged?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the
github-actions[bot] closed pull request #41628: [SPARK-38230][SQL]
InsertIntoHadoopFsRelationCommand unnecessarily fetches details of partitions
URL: https://github.com/apache/spark/pull/41628
--
This is an automated message from the Apache Git Service.
To respond to the message, please log
github-actions[bot] commented on PR #41628:
URL: https://github.com/apache/spark/pull/41628#issuecomment-1806954903
We're closing this PR because it hasn't been updated in a while. This isn't
a judgement on the merit of the PR in any way. It's just a way of keeping the
PR queue manageable.