pg.ddx.io  pgsql-committers@postgresql.org mailing list archive  
help / color / mirror / Atom feed
pgsql: Ignore extended statistics for inheritance trees
6+ messages / 1 participants
[nested] [flat]

* pgsql: Ignore extended statistics for inheritance trees
@ 2022-01-15 02:10 Tomas Vondra <tomas.vondra@postgresql.org>
  0 siblings, 0 replies; 6+ messages in thread

From: Tomas Vondra @ 2022-01-15 02:10 UTC (permalink / raw)
  To: pgsql-committers@lists.postgresql.org

Ignore extended statistics for inheritance trees

Since commit 859b3003de we only build extended statistics for individual
relations, ignoring the child relations. This resolved the issue with
updating catalog tuple twice, but we still tried to use the statistics
when calculating estimates for the whole inheritance tree. When the
relations contain very distinct data, it may produce bogus estimates.

This is roughly the same issue 427c6b5b9 addressed ~15 years ago, and we
fix it the same way - by ignoring extended statistics when calculating
estimates for the inheritance tree as a whole. We still consider
extended statistics when calculating estimates for individual child
relations, of course.

This may result in plan changes due to different estimates, but if the
old statistics were not describing the inheritance tree particularly
well it's quite likely the new plans is actually better.

Report and patch by Justin Pryzby, minor fixes and cleanup by me.
Backpatch all the way back to PostgreSQL 10, where extended statistics
were introduced (same as 859b3003de).

Author: Justin Pryzby
Reported-by: Justin Pryzby
Backpatch-through: 10
Discussion: https://postgr.es/m/20210923212624.GI831%40telsasoft.com

Branch
------
master

Details
-------
https://git.postgresql.org/pg/commitdiff/36c4bc6e725fe3e9b5a6cc7f1f83010534f18e4b

Modified Files
--------------
src/backend/statistics/dependencies.c   |  9 ++++++++
src/backend/statistics/extended_stats.c |  9 ++++++++
src/backend/utils/adt/selfuncs.c        | 17 ++++++++++++++
src/test/regress/expected/stats_ext.out | 41 +++++++++++++++++++++++++++++++++
src/test/regress/sql/stats_ext.sql      | 22 ++++++++++++++++++
5 files changed, 98 insertions(+)



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* pgsql: Ignore extended statistics for inheritance trees
@ 2022-01-15 02:12 Tomas Vondra <tomas.vondra@postgresql.org>
  0 siblings, 0 replies; 6+ messages in thread

From: Tomas Vondra @ 2022-01-15 02:12 UTC (permalink / raw)
  To: pgsql-committers@lists.postgresql.org

Ignore extended statistics for inheritance trees

Since commit 859b3003de we only build extended statistics for individual
relations, ignoring the child relations. This resolved the issue with
updating catalog tuple twice, but we still tried to use the statistics
when calculating estimates for the whole inheritance tree. When the
relations contain very distinct data, it may produce bogus estimates.

This is roughly the same issue 427c6b5b9 addressed ~15 years ago, and we
fix it the same way - by ignoring extended statistics when calculating
estimates for the inheritance tree as a whole. We still consider
extended statistics when calculating estimates for individual child
relations, of course.

This may result in plan changes due to different estimates, but if the
old statistics were not describing the inheritance tree particularly
well it's quite likely the new plans is actually better.

Report and patch by Justin Pryzby, minor fixes and cleanup by me.
Backpatch all the way back to PostgreSQL 10, where extended statistics
were introduced (same as 859b3003de).

Author: Justin Pryzby
Reported-by: Justin Pryzby
Backpatch-through: 10
Discussion: https://postgr.es/m/20210923212624.GI831%40telsasoft.com

Branch
------
REL_14_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/2cc007fd0359870b4a1e959af0ac94e61df47240

Modified Files
--------------
src/backend/statistics/dependencies.c   |  9 ++++++++
src/backend/statistics/extended_stats.c |  9 ++++++++
src/backend/utils/adt/selfuncs.c        | 17 ++++++++++++++
src/test/regress/expected/stats_ext.out | 41 +++++++++++++++++++++++++++++++++
src/test/regress/sql/stats_ext.sql      | 22 ++++++++++++++++++
5 files changed, 98 insertions(+)



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* pgsql: Ignore extended statistics for inheritance trees
@ 2022-01-15 02:13 Tomas Vondra <tomas.vondra@postgresql.org>
  0 siblings, 0 replies; 6+ messages in thread

From: Tomas Vondra @ 2022-01-15 02:13 UTC (permalink / raw)
  To: pgsql-committers@lists.postgresql.org

Ignore extended statistics for inheritance trees

Since commit 859b3003de we only build extended statistics for individual
relations, ignoring the child relations. This resolved the issue with
updating catalog tuple twice, but we still tried to use the statistics
when calculating estimates for the whole inheritance tree. When the
relations contain very distinct data, it may produce bogus estimates.

This is roughly the same issue 427c6b5b9 addressed ~15 years ago, and we
fix it the same way - by ignoring extended statistics when calculating
estimates for the inheritance tree as a whole. We still consider
extended statistics when calculating estimates for individual child
relations, of course.

This may result in plan changes due to different estimates, but if the
old statistics were not describing the inheritance tree particularly
well it's quite likely the new plans is actually better.

Report and patch by Justin Pryzby, minor fixes and cleanup by me.
Backpatch all the way back to PostgreSQL 10, where extended statistics
were introduced (same as 859b3003de).

Author: Justin Pryzby
Reported-by: Justin Pryzby
Backpatch-through: 10
Discussion: https://postgr.es/m/20210923212624.GI831%40telsasoft.com

Branch
------
REL_13_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/acfde7c5837dd3c2a40ff9cc639ba5d914e8e8ea

Modified Files
--------------
src/backend/statistics/dependencies.c   |  9 ++++++++
src/backend/statistics/extended_stats.c |  9 ++++++++
src/backend/utils/adt/selfuncs.c        |  8 +++++++
src/test/regress/expected/stats_ext.out | 41 +++++++++++++++++++++++++++++++++
src/test/regress/sql/stats_ext.sql      | 22 ++++++++++++++++++
5 files changed, 89 insertions(+)



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* pgsql: Ignore extended statistics for inheritance trees
@ 2022-01-15 02:16 Tomas Vondra <tomas.vondra@postgresql.org>
  0 siblings, 0 replies; 6+ messages in thread

From: Tomas Vondra @ 2022-01-15 02:16 UTC (permalink / raw)
  To: pgsql-committers@lists.postgresql.org

Ignore extended statistics for inheritance trees

Since commit 859b3003de we only build extended statistics for individual
relations, ignoring the child relations. This resolved the issue with
updating catalog tuple twice, but we still tried to use the statistics
when calculating estimates for the whole inheritance tree. When the
relations contain very distinct data, it may produce bogus estimates.

This is roughly the same issue 427c6b5b9 addressed ~15 years ago, and we
fix it the same way - by ignoring extended statistics when calculating
estimates for the inheritance tree as a whole. We still consider
extended statistics when calculating estimates for individual child
relations, of course.

This may result in plan changes due to different estimates, but if the
old statistics were not describing the inheritance tree particularly
well it's quite likely the new plans is actually better.

Report and patch by Justin Pryzby, minor fixes and cleanup by me.
Backpatch all the way back to PostgreSQL 10, where extended statistics
were introduced (same as 859b3003de).

Author: Justin Pryzby
Reported-by: Justin Pryzby
Backpatch-through: 10
Discussion: https://postgr.es/m/20210923212624.GI831%40telsasoft.com

Branch
------
REL_12_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/76569ad6f423f56f655699f57c39298af2111914

Modified Files
--------------
src/backend/statistics/dependencies.c   |  9 ++++++++
src/backend/statistics/extended_stats.c |  9 ++++++++
src/backend/utils/adt/selfuncs.c        |  8 +++++++
src/test/regress/expected/stats_ext.out | 41 +++++++++++++++++++++++++++++++++
src/test/regress/sql/stats_ext.sql      | 22 ++++++++++++++++++
5 files changed, 89 insertions(+)



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* pgsql: Ignore extended statistics for inheritance trees
@ 2022-01-15 02:17 Tomas Vondra <tomas.vondra@postgresql.org>
  0 siblings, 0 replies; 6+ messages in thread

From: Tomas Vondra @ 2022-01-15 02:17 UTC (permalink / raw)
  To: pgsql-committers@lists.postgresql.org

Ignore extended statistics for inheritance trees

Since commit 859b3003de we only build extended statistics for individual
relations, ignoring the child relations. This resolved the issue with
updating catalog tuple twice, but we still tried to use the statistics
when calculating estimates for the whole inheritance tree. When the
relations contain very distinct data, it may produce bogus estimates.

This is roughly the same issue 427c6b5b9 addressed ~15 years ago, and we
fix it the same way - by ignoring extended statistics when calculating
estimates for the inheritance tree as a whole. We still consider
extended statistics when calculating estimates for individual child
relations, of course.

This may result in plan changes due to different estimates, but if the
old statistics were not describing the inheritance tree particularly
well it's quite likely the new plans is actually better.

Report and patch by Justin Pryzby, minor fixes and cleanup by me.
Backpatch all the way back to PostgreSQL 10, where extended statistics
were introduced (same as 859b3003de).

Author: Justin Pryzby
Reported-by: Justin Pryzby
Backpatch-through: 10
Discussion: https://postgr.es/m/20210923212624.GI831%40telsasoft.com

Branch
------
REL_11_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/b3cac25f4d50d66ef08a9a706df045c888578f8f

Modified Files
--------------
src/backend/statistics/dependencies.c   |  9 +++++
src/backend/statistics/extended_stats.c |  1 +
src/backend/utils/adt/selfuncs.c        |  8 +++++
src/test/regress/expected/stats_ext.out | 61 +++++++++++++++++++++++++++++++++
src/test/regress/sql/stats_ext.sql      | 43 +++++++++++++++++++++++
5 files changed, 122 insertions(+)



^ permalink  raw  reply  [nested|flat] 6+ messages in thread

* pgsql: Ignore extended statistics for inheritance trees
@ 2022-01-15 02:20 Tomas Vondra <tomas.vondra@postgresql.org>
  0 siblings, 0 replies; 6+ messages in thread

From: Tomas Vondra @ 2022-01-15 02:20 UTC (permalink / raw)
  To: pgsql-committers@lists.postgresql.org

Ignore extended statistics for inheritance trees

Since commit 859b3003de we only build extended statistics for individual
relations, ignoring the child relations. This resolved the issue with
updating catalog tuple twice, but we still tried to use the statistics
when calculating estimates for the whole inheritance tree. When the
relations contain very distinct data, it may produce bogus estimates.

This is roughly the same issue 427c6b5b9 addressed ~15 years ago, and we
fix it the same way - by ignoring extended statistics when calculating
estimates for the inheritance tree as a whole. We still consider
extended statistics when calculating estimates for individual child
relations, of course.

This may result in plan changes due to different estimates, but if the
old statistics were not describing the inheritance tree particularly
well it's quite likely the new plans is actually better.

Report and patch by Justin Pryzby, minor fixes and cleanup by me.
Backpatch all the way back to PostgreSQL 10, where extended statistics
were introduced (same as 859b3003de).

Author: Justin Pryzby
Reported-by: Justin Pryzby
Backpatch-through: 10
Discussion: https://postgr.es/m/20210923212624.GI831%40telsasoft.com

Branch
------
REL_10_STABLE

Details
-------
https://git.postgresql.org/pg/commitdiff/ff0e7c7e840122419054d72a7884fa17ce1a4ab9

Modified Files
--------------
src/backend/statistics/dependencies.c   |  9 +++++
src/backend/statistics/extended_stats.c |  1 +
src/backend/utils/adt/selfuncs.c        |  8 ++++
src/test/regress/expected/stats_ext.out | 65 +++++++++++++++++++++++++++++++++
src/test/regress/sql/stats_ext.sql      | 47 ++++++++++++++++++++++++
5 files changed, 130 insertions(+)



^ permalink  raw  reply  [nested|flat] 6+ messages in thread


end of thread, other threads:[~2022-01-15 02:20 UTC | newest]

Thread overview: 6+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2022-01-15 02:10 pgsql: Ignore extended statistics for inheritance trees Tomas Vondra <tomas.vondra@postgresql.org>
2022-01-15 02:12 pgsql: Ignore extended statistics for inheritance trees Tomas Vondra <tomas.vondra@postgresql.org>
2022-01-15 02:13 pgsql: Ignore extended statistics for inheritance trees Tomas Vondra <tomas.vondra@postgresql.org>
2022-01-15 02:16 pgsql: Ignore extended statistics for inheritance trees Tomas Vondra <tomas.vondra@postgresql.org>
2022-01-15 02:17 pgsql: Ignore extended statistics for inheritance trees Tomas Vondra <tomas.vondra@postgresql.org>
2022-01-15 02:20 pgsql: Ignore extended statistics for inheritance trees Tomas Vondra <tomas.vondra@postgresql.org>

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox