Zoltan Borok-Nagy has posted comments on this change. (
http://gerrit.cloudera.org:8080/24394 )
Change subject: IMPALA-15065: Fix incorrect partition spec identification when
removing deletion vectors from Iceberg V3 tables
......................................................................
Patch Set 6: Code-Review+2
(1 comment)
Tested the change on the "Extended One Trillion Row Challenge", scaled down to
10 Billion records and a single machine.
The change can introduce some overhead in "Executed Iceberg operation", but it
seems tolerable at this scale. The manifest files were read locally, so in a
real environment the slowness can be higher, especially when the manifest cache
cannot contain every entry. Once Iceberg fixes the underlying issue we can
re-consider the original approach.
"Executed Iceberg operation" time in profile:
Base This
- DML_1: 133ms vs 401ms (+268ms)
- DML_2: 252ms vs 325ms (+73ms)
- DML_3: 167ms vs 178ms (+12ms)
- DML_4: 153ms vs 137ms (−16ms)
- DML_5: 151ms vs 197ms (+46ms)
- DML_6: 204ms vs 282ms (+78ms)
http://gerrit.cloudera.org:8080/#/c/24394/6//COMMIT_MSG
Commit Message:
http://gerrit.cloudera.org:8080/#/c/24394/6//COMMIT_MSG@8
PS6, Line 8: d
I think title should remain a single line
--
To view, visit http://gerrit.cloudera.org:8080/24394
To unsubscribe, visit http://gerrit.cloudera.org:8080/settings
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-MessageType: comment
Gerrit-Change-Id: Iba280972e2089e92771c8b3f7355d916d46ad968
Gerrit-Change-Number: 24394
Gerrit-PatchSet: 6
Gerrit-Owner: Peter Rozsa <[email protected]>
Gerrit-Reviewer: Impala Public Jenkins <[email protected]>
Gerrit-Reviewer: Peter Rozsa <[email protected]>
Gerrit-Reviewer: Zoltan Borok-Nagy <[email protected]>
Gerrit-Comment-Date: Thu, 11 Jun 2026 11:51:48 +0000
Gerrit-HasComments: Yes
