Mar 2022

1 commit
Invent recursive_worktable_factor GUC to replace hard-wired constant.
24 Mar 2022Tom Lanemaster15beta1REL_19_STABLE15beta1REL_18_STABLE15beta1REL_17_STABLE15beta1REL_16_STABLE15beta1REL_15_STABLE15beta1
LLM analysis, 100% suremodified before commitmoderate changesfrom patchset #121 of 4 reviewers saw the final version

The commit retains the six-file implementation and clamp_row_est() from #12, renaming the setting to recursive_worktable_factor as proposed in #13. It removes the zero-as-default branch in src/backend/optimizer/path/costsize.c, raises the GUC minimum to 0.001 in src/backend/utils/misc/guc.c, and revises documentation and comments.

  • Kenaniah Cernysaw an earlier versionLLM analysis, 94% sureIn #3, he examined the supplied EXPLAIN estimates, warned about nonlinear estimation, and endorsed the configurable multiplier in #1; row clamping and bounds changes followed.
  • Peter Eisentrautsaw an earlier versionLLM analysis, 96% sureIn #4, he accepted the setting in #1 as a practical compromise and endorsed its graph OLTP versus OLAP rationale; row clamping and bounds changes followed.
  • Robert Haassaw an earlier versionLLM analysis, 93% sureIn #6, he endorsed the proposal in #1 as harmless and practical given the difficulty of better automatic estimates; row clamping and bounds changes followed.
  • Tom Lanesaw the final versioncommittercommitted the patch

Credits across 1 commit

Author
Simon RiggsLLM1
Reviewer
Tom Lanecommitter1
Committer
Tom Lane1

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