Apr 2019

1 commit
Clean up handling of constraint_exclusion and enable_partition_pruning.
30 Apr 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+2 moreAlso in: selecting from partitions and constraint exclusion (not this thread)
LLM analysis, 100% surelanded as postedfrom patchset #2631 of 1 reviewer saw the final version

The canonical commit exactly matches the HEAD attachment in #263, including plancat.c policy flags, regression tests, and documentation; the other attachment is a branch alternative. This submission is newer than the March sources in thread #40312, and the commit credits Amit Langote and sender Tom Lane jointly.

  • Tom Lanesaw the final versioncommittercommitted the patch

Mar 2019

8 commits
Add test case exercising formerly-unreached code in inheritance_planner.
31 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surewritten from the discussion, no matching patch

The unattached updatable-view UPDATE testcase follows Amit's challenge to target-index remapping in #224 and Tom's explanation and request for a reproducer in #225. No submitted patch contains this testcase.

Speed up planning when partitions can be pruned at plan time.
30 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #2281 of 7 reviewers saw the final version

The second patch in #228 reproduces the commit except that partprune.c's subpart_map memset uses sizeof(int) instead of sizeof(Oid). The GEQO prerequisite landed separately; the commit credits Amit despite Tom posting the final submission.

  • Dilip Kumarsaw an earlier versionLLM analysis, 97% sureReviewed early delayed-expansion code in #10 and #20, and the commit explicitly credits him; substantial planner redesign followed.
  • Jesper Pedersensaw an earlier versionLLM analysis, 99% sureReported benchmarks, profiling, check-world results, and a postgres_fdw failure, most recently testing #199 in #200 before Tom's inheritance-planner rework.
  • Imai, Yoshikazusaw an earlier versionLLM analysis, 98% sureSupplied extensive code review and memory/performance tests, including hash-partition UPDATE benchmarks of #221 in #223, before the partitionwise-join redesign in #228.
  • David Rowleysaw an earlier versionLLM analysis, 99% sureIdentified locking and equivalence-class problems in #15, supplied diagnostics in #39, and reviewed later revisions before the major redesign.
  • Justin Pryzbysaw an earlier versionLLM analysis, 98% sureTested a large customer schema in #52 and caught the incorrect IS_SIMPLE_REL operand in #143, which was corrected before subsequent architectural changes.
  • Kato, Shosaw an earlier versionLLM analysis, 95% sureBenchmarked the initial submitted implementation in #8, before the extensive later reworking of partition expansion and target planning.
  • Tom Lanesaw the final versioncommittercommitted the patch
Avoid crash in partitionwise join planning under GEQO.
30 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+2 more
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #2281 of 1 reviewer saw the final version

The first patch in #228 supplies the committed GEQO and partitionwise-join implementation; final differences update regression-test comments. Amit requested those corrections in #229, and Tom accepted them in #230.

  • Amit Langotesaw the final versionLLM analysis, 100% sureIdentified regression-test comments invalidated by the new join plans in #229, and Tom incorporated the corrections described in #230.
Avoid passing query tlist around separately from root->processed_tlist.
27 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surewritten from the discussion, no matching patch

Tom's removal of targetlist aliases and parameters implements the processed_tlist concerns in #202 and #203, followed by Amit's adoption of direct updates in #205. No attachment contains this standalone cleanup.

Build "other rels" of appendrel baserels in a separate step.
26 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surepart of a posted patchsetmoderate changesfrom patchset #2061 of 5 reviewers saw the final version

Extracts the first member of #206, separating otherrel construction from baserel creation. Tom additionally moves lateral-reference propagation into build_simple_rel and simplifies create_lateral_join_info, as explained in #209.

  • Dilip Kumarsaw an earlier versionLLM analysis, 94% sureReviewed the earlier delayed relation-construction work in #20 and is explicitly credited in the commit, before its extraction and later restructuring.
  • Jesper Pedersensaw an earlier versionLLM analysis, 98% sureTested the series containing separate otherrel construction and supplied adopted corrections in #152, before the final lateral-reference rewrite.
  • Imai, Yoshikazusaw an earlier versionLLM analysis, 100% sureExplicitly reviewed the first patch of #206 in #207, confirming its changes and finding a typo before Tom rewrote lateral-information propagation.
  • David Rowleysaw an earlier versionLLM analysis, 94% sureReviewed the earlier lazy relation-construction implementation in #15 and #69, and the commit explicitly credits him despite substantial subsequent restructuring.
  • Tom Lanesaw the final versioncommittercommitted the patch
Get rid of duplicate child RTE for a partitioned table.
26 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 100% surepart of a posted patchsetmoderate changesfrom patchset #2061 of 1 reviewer saw the final version

Extracts duplicate partition-parent RTE removal from #206. Tom expands the five-line deletion with local simplification in expand_single_inheritance_child, comments, and postgres_fdw expected-output repairs, explicitly described in #208.

  • Tom Lanesaw the final versioncommittercommitted the patch
Rearrange make_partitionedrel_pruneinfo to avoid work when we can't prune.
22 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #1992 of 2 reviewers saw the final version

Extracts #199's make_partitionedrel_pruneinfo optimization, retaining its two-phase construction and merged mapping loop. Changes are primarily comments, harmless statement ordering, and freeing the temporary map on the early-return path.

  • Jesper Pedersensaw the final versionLLM analysis, 100% sureConfirmed that #199 removed the partition-count-dependent throughput loss and passed check-world in #200.
  • Tom Lanesaw the final versioncommittercommitted the patch
Don't copy PartitionBoundInfo in set_relation_partition_info.
22 Mar 2019Tom Lanemaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 moreAlso in: hyrax vs. RelationBuildPartitionDesc (not this thread)
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #1993 of 3 reviewers saw the final version

Extracts #199's direct PartitionBoundInfo assignment in plancat.c, removing an unused include and series-dependent guard/comment context. thread #40272 supplies related lifetime discussion, but its recorded verdict also identifies this thread as the implementation source.

  • Imai, Yoshikazusaw the final versionLLM analysis, 99% sureBenchmarked the avoid-copy member of #183 separately in #186, demonstrating its large incremental throughput improvement.
  • Jesper Pedersensaw the final versionLLM analysis, 96% sureTested the complete series containing this optimization, including a passing check-world result for #199 in #200.
  • Tom Lanesaw the final versioncommittercommitted the patch

Feb 2019

1 commit
Move building of child base quals out into a new function
1 Feb 2019Alvaro Herreramaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 100% suremodified before commitlight changesfrom patchset #1081 of 1 reviewer saw the final version

Matches David's standalone apply_child_basequals extraction in #108, with comment and formatting edits. Alvaro explicitly reports committing it with amended comments in #109.

  • Alvaro Herrerasaw the final versioncommittercommitted the patch

Jan 2019

2 commits
Reorganize planner code moved in b60c39759908
16 Jan 2019Alvaro Herreramaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #761 of 1 reviewer saw the final version

Takes the fixup member most recently submitted in #76, originally developed in #68 and #71. Alvaro additionally reorders appendrel helper functions and declarations, confirmed in #77.

  • Alvaro Herrerasaw the final versioncommittercommitted the patch
Move inheritance expansion code into its own file
10 Jan 2019Alvaro Herreramaster12beta1REL_19_STABLE12beta1REL_18_STABLE12beta1REL_17_STABLE12beta1REL_16_STABLE12beta1REL_15_STABLE12beta1REL_14_STABLE12beta1REL_13_STABLE12beta1+1 more
LLM analysis, 100% surepart of a posted patchsetlight changesfrom patchset #633 of 3 reviewers saw the final version

Commits only the first member of #63, moving inheritance expansion and appendrel-expression helpers into inherit.c and appendinfo.c. Differences are include cleanup and formatting, matching the announcement in #65.

  • Jesper Pedersensaw the final versionLLM analysis, 88% sureRan check-world and benchmarks on #37's series containing the separated code-movement patch, reporting success in #38; its later changes were reordering and mechanical cleanup.
  • Justin Pryzbysaw the final versionLLM analysis, 86% sureTested #51's complete series, including the code movement, against a large customer schema in #52 before the movement was reordered for independent commitment.
  • Alvaro Herrerasaw the final versioncommittercommitted the patch

Credits across 12 commits

Author
Amit Langote7 LLM8
David Rowley1
Tom LaneLLM1
Reviewer
Tom Lanecommitter6
Alvaro Herreracommitter3
Committer
Tom Lane9
Alvaro Herrera3

N LLM means N of the credits come from LLM analysis, not from commit trailers.