HIVE-26832: Implement SHOW PARTITIONS for Iceberg Tables - #4346
Conversation
| throw new HiveException(ErrorMsg.TABLE_NOT_PARTITIONED, desc.getTabName()); | ||
| boolean isIcebergTable = isIcebergTable(tbl); | ||
| if (!(tbl.isPartitioned() || isIcebergTable)) { | ||
| throw new HiveException(ErrorMsg.TABLE_NOT_PARTITIONED, desc.getTabName()); |
There was a problem hiding this comment.
Can we generalize this part rather than hardcoding ICEBERG? I guess this is useful for another storage handler.
There was a problem hiding this comment.
Haha. We've been developing a tailored storage with some similar concepts to Iceberg for 10 years. We really like recent changes for Iceberg because it would give us a chance to move our hacks to standard APIs of storage handlers. Let me check newly added APIs and propose better generalization.
There was a problem hiding this comment.
Haha. We've been developing a tailored storage with some similar concepts to Iceberg for 10 years. We really like recent changes for Iceberg because it would give us a chance to move our hacks to standard APIs of storage handlers. Let me check newly added APIs and propose better generalization.
Yes, Iceberg has a very elegant API and solutions. I like it, also the concept removes lot of hacks from the hive side as well.
There was a problem hiding this comment.
Can we generalize this part rather than hardcoding ICEBERG? I guess this is useful for another storage handler.
Done.
There was a problem hiding this comment.
this block is to handle cases when you call show partitions on non-partitioned table. no need to protect if there is no partition support, especially when you added the default impl for showPartitions in storage handler
|
Kudos, SonarCloud Quality Gate passed! |








What changes were proposed in this pull request?
SHOW PARTITIONS support for iceberg tables.
Why are the changes needed?
For Hive partition table , the show partition command outputs the following:
In this PR, we are implementing the same feature for hive iceberg tables:
With partition evolution, the output will be per spec id:
Does this PR introduce any user-facing change?
How was this patch tested?
Q test