OSDN Git Service
Kyotaro Horiguchi [Thu, 18 May 2017 09:43:09 +0000 (18:43 +0900)]
Fix a crash bug on complex views when enable_hint_table is on
The Query that planner receives sometimes irrelevant to
debug_query_string. If enable_hint_table is on, pg_hint_plan_planner
normalizes debug_query_string using query-jumble information created
from the irrelevant Query the can lead to crash. To avoid this
situation, retrieve hints in post_parse_analyze_hook, where
corresponding pairs of a query string and a parsed Query.
Kyotaro Horiguchi [Thu, 18 May 2017 01:40:48 +0000 (10:40 +0900)]
Follow the behavioral change of parallel execution
The behavior of parallel planning is changed until 9.6.3, let
expected/ut-W.out to follow that.
Kyotaro Horiguchi [Thu, 18 May 2017 01:28:49 +0000 (10:28 +0900)]
Fix a bug of Parallel hint
During manipulation on pathlist of a GatherPath, foreach was mitakenly
used for a loop using list_delete_cell. This leads to a crash with
SEGV.
Kyotaro Horiguchi [Tue, 21 Feb 2017 00:27:55 +0000 (09:27 +0900)]
Provide upgrade path from 1.1.3
pg_hint_plan96-1.2.0 has an upgrade path from 1.1.3. This adds
requires installation sql file and puts pointers to the files in
Makefile and spec file. It is the only upgrade path of pg_hint_plan
1.2.0 since it works only on PG9.6 and versions earlier than 1.1.3
don't support PG9.6.
Kyotaro Horiguchi [Thu, 16 Feb 2017 10:08:54 +0000 (19:08 +0900)]
Don't apply parallel hint on rels that planner considers unsafe.
pg_hint_plan considered had its own decision logic on whether
parallel-safe or not but it is dangerous and unnecessary. Just follow
planner's decision.
Kyotaro Horiguchi [Thu, 16 Feb 2017 08:41:30 +0000 (17:41 +0900)]
Reconsider hinting process
I rethinked the steps of enforcement considering the parallel planning
code. pg_hint_plan_set_rel_pathlist() gets simpler and easier to read
(maybe).
As the result of this patch, some behaviors has been changed.
- False generation of gahter path caused by hinting on the other side
of joins.
- Scan hints work with Parallel hints. (It's useless for the moment,
though).
Tatsuro Yamada [Thu, 16 Feb 2017 02:00:32 +0000 (11:00 +0900)]
Fix a typo of documentation.
Author: Tatsuro Yamada <yamatattsu@gmail.com>
Kyotaro Horiguchi [Mon, 13 Feb 2017 08:01:49 +0000 (17:01 +0900)]
Don't set_plain_rel_pathlist for other than RTE_RELATION.
pg_hint_plan_set_rel_pathlist can call set_plain_rel_pathlist for
relations other than RTE_RELATION and this results in a crash.
Kyotaro Horiguchi [Mon, 13 Feb 2017 07:56:44 +0000 (16:56 +0900)]
Following a fix in PG9.6.1.
PG9.6.2 gets a fix on planning that affects on the result of
ut-S. ut-S follows that by this fix. As the result ut-S won't success
for 9.6.1 or ealier after this patch.
Kyotaro Horiguchi [Mon, 13 Feb 2017 07:52:45 +0000 (16:52 +0900)]
Fix parallel hint error handling.
Some kind of error of a parallel hint resulted confusing log
output. Addition to that, due to improper assumption of .conf file,
ordinary setup of max_worker_processes caused failure of ut-W. Fixed
them.
Kyotaro Horiguchi [Fri, 10 Feb 2017 08:19:43 +0000 (17:19 +0900)]
Refactor hint application mechamism next step
I removed useless code forgotten to be removed in
add_paths_to_joinrel_wrapper by the previous refactoring. Addition to
that, find_scan_method_hint and find_parallel_hint are found not to
require the third parameter rel, so removed it. Some inadequate tests
are fixed and changed according to change of unclear behaviors.
Kyotaro Horiguchi [Fri, 10 Feb 2017 02:06:05 +0000 (11:06 +0900)]
Improve error handling of Parallel hint
Improve poor error report for Parallel hint.
Kyotaro Horiguchi [Fri, 10 Feb 2017 02:09:52 +0000 (11:09 +0900)]
Fix wrongly lowering max_parallel_workers_per_gather
Number of parallel workers of non-hinted relations is wrongly forced
to be zero even if max_parallel_workers_per_gather is more than
zero. Non-hinted relations get to have max_parallel_workes_per_gather
as set by this fix.
Kyotaro Horiguchi [Thu, 9 Feb 2017 01:50:29 +0000 (10:50 +0900)]
Refactor hint application mechamism
Before we have the new hook set_rel_pathlist_hook, plan enforcment is
performed in two steps, one works only for queries with one relation,
the other works only for joins. Now they are performed once and at
once in set_rel_pathlist_hook. As a byproduct of this refactoring,
index restriction no longer runs for indexes on constraint-excluded
relations and expected/ut-S.out changed a bit.
The major changes by this patch are the followings.
- Remove hook to get_relation_info_hook:
The function for the hook pg_hint_plan_get_relation_info is removed.
- pg_hint_plan_join_search no longer calls rebuild_scan_path. The
function and a copied function set_plain_rel_pathlist are removed.
- delete_indexes has been renamed to restrict_indexes, which
represents the functionality clearer. And the function was
refactored main for readablity.
- Added several regtests for parallel.
Kyotaro Horiguchi [Thu, 9 Feb 2017 00:07:14 +0000 (09:07 +0900)]
Fix default_version of extension control file
default_version in pg_hint_plan.control was left unchanged on updating
version number. Fixed it.
Kyotaro Horiguchi [Thu, 9 Feb 2017 00:06:28 +0000 (09:06 +0900)]
Edit copyright year
Kyotaro Horiguchi [Wed, 8 Feb 2017 01:34:09 +0000 (10:34 +0900)]
Properly change the state of parallel hints.
The state member of parallel hint is forgotten to be chaned on usage.
This fixes it.
Kyotaro Horiguchi [Wed, 8 Feb 2017 01:16:01 +0000 (10:16 +0900)]
Rename some functions and refactor code.
Some functions had a bit confusing naming and inconvenient interfaces.
Renamed such functions and refactored the code.
Kyotaro Horiguchi [Tue, 7 Feb 2017 11:28:18 +0000 (20:28 +0900)]
Change version to 1.2.0.
Change version number to 1.2.0 and prepare RPM build envoronment.
Kyotaro Horiguchi [Tue, 7 Feb 2017 11:05:12 +0000 (20:05 +0900)]
Add description of Parallel hint to documentation
Addition to that, documents are refactored.
Kyotaro Horiguchi [Tue, 7 Feb 2017 11:03:13 +0000 (20:03 +0900)]
New hint "Parallel"
Added a new feature to control parallel query execution.
Kyotaro Horiguchi [Tue, 7 Feb 2017 03:54:05 +0000 (12:54 +0900)]
Rename current_hint to current_hint_state
The name current_hint is quite confusing so change the name to less
confusing one.
Kyotaro Horiguchi [Mon, 6 Feb 2017 06:18:08 +0000 (15:18 +0900)]
Fix some syntax and comments.
Some following comma for the last element in enum definition and
incorrect comments are fixed.
Kyotaro Horiguchi [Mon, 5 Sep 2016 07:18:13 +0000 (16:18 +0900)]
Remove an unnecessary Makefile entry for tar file
Remove mistakenly copy'n-pasted description from outside in Makefile.
Kyotaro Horiguchi [Mon, 5 Sep 2016 07:11:30 +0000 (16:11 +0900)]
Edit copyright notices
Kyotaro Horiguchi [Thu, 1 Sep 2016 04:57:30 +0000 (13:57 +0900)]
Support PostgreSQL 9.6 RC1
Edited Makefile and renamed and edited RPM spec file.
Copied functions are updated and follow the changes of internal API.
Modified regressions to pass with 9.6 RC1.
Kyotaro Horiguchi [Fri, 13 May 2016 06:42:45 +0000 (15:42 +0900)]
Add RPM spec file forgotten in the previous commit.
Kyotaro HORIGUCHI [Tue, 26 Apr 2016 08:44:48 +0000 (17:44 +0900)]
Merge pull request #36272 (fujii_masao/pg_hint_plan/master into master).
MasaoFujii [Fri, 22 Apr 2016 04:38:42 +0000 (13:38 +0900)]
Use pg_reload_conf() to reload the configuration file in regression test.
Previously the regression test ran pg_ctl reload command for that purpose.
However in the environment where the path to PostgreSQL binaries is not set,
the regression test always failed because pg_ctl could not be found.
To reload the configuration file even in that environment, this commit
changes the regression test so that it uses pg_reload_conf function, instead.
Kyotaro Horiguchi [Fri, 15 Jan 2016 02:48:52 +0000 (11:48 +0900)]
Prepare for 1.1.3/PG95.
The pg_hint_plan version of master branch is left as 1.1.2. This
commit corrects the pg_hint_plan version, and edits RPM building
environment for PostgreSQL 9.5.
Kyotaro Horiguchi [Fri, 15 Jan 2016 02:47:17 +0000 (11:47 +0900)]
Edit copyright years.
Kyotaro Horiguchi [Fri, 15 Jan 2016 02:44:37 +0000 (11:44 +0900)]
Change for new syntax items in 9.5.
Some syntactical elements are added in 9.5. This change reflects it.
Addition to that, deleted unused parts in pg_stat_statements, which
are previously invalidated with #ifdef directives.
Kyotaro Horiguchi [Fri, 15 Jan 2016 00:27:24 +0000 (09:27 +0900)]
Follow a change of error message.
Context message of PL/pgSQL is changed so this commit refrects it to
exted/pg_hint_plan.out.
Kyotaro Horiguchi [Fri, 15 Jan 2016 00:28:29 +0000 (09:28 +0900)]
Followed the changes of 9.5.0 release.
There was some changes in 9.5.0 release affect
pg_hint_plan. set_append_rel_pathlist() no longer sets cheapest path
and it became a business of the caller. So rebuild_scan_path() does
so. core.c gets changed from changing lateral join infrastracture
(
acfcd45cacb6df23edba4cb3753a2be594238a99) and a change related to custom path(
c2ea2285e978d9289084846a3343cef7d261d880).
Kyotaro Horiguchi [Mon, 17 Aug 2015 09:33:38 +0000 (18:33 +0900)]
Made compatible with 9.5a1
Make make_join_rel.c to catch up to REL9_5_ALPHA2. Change regtests to
reflect some differences of PostgreSQL's behavior in such like
collapsing of simple VALUES clause, rows estimation and explain
representation of join filter.
Kyotaro Horiguchi [Mon, 22 Dec 2014 08:34:15 +0000 (17:34 +0900)]
Fixed a bug related to SQL statements in PL/pgSQL functions.
At least since 9.1, true query strings of SQL statements executed in
PL/pgSQL functions were found that were not obtained correctly by the
previous implement, it is because that PLpgSQL_stmt variable did not
have proper query string to be read for hints. Instead, it is changed
to read them from the top of error_context_stack in
pg_hint_plan_planner(). This change made a slight difference in
behavior which doesn't affect its work so a part of the regtest was
also changed. And added some regression tests for it.
Kyotaro Horiguchi [Mon, 22 Dec 2014 02:07:09 +0000 (11:07 +0900)]
Catch up the changes of pg_stat_statements.c
The change has nothing to do with pg_hint_plan but it's necessary to
catch it up in order to keep the difference minimal for the ease of
maintenance.
Kyotaro Horiguchi [Tue, 16 Dec 2014 10:00:50 +0000 (19:00 +0900)]
Set additional parameter of set_config_option().
9.5dev added new parameter to set_config_option(). This paramter
should be false in set_config_option_wrapper.
Kyotaro Horiguchi [Tue, 16 Dec 2014 10:12:59 +0000 (19:12 +0900)]
Fix regression test according to the changes of PostgreSQL.
PostgreSQL became to give WARNING for usage of hash indexes. This
change is to follow it.
Kyotaro Horiguchi [Wed, 17 Dec 2014 02:20:21 +0000 (11:20 +0900)]
Prepare for 1.1.2
Kyotaro Horiguchi [Fri, 19 Dec 2014 07:10:46 +0000 (16:10 +0900)]
Remove unnecessary tracker code from documentations.
Kyotaro Horiguchi [Fri, 19 Dec 2014 06:43:36 +0000 (15:43 +0900)]
Added English documentation.
Kyotaro Horiguchi [Fri, 19 Dec 2014 06:05:58 +0000 (15:05 +0900)]
Some fix and addition to the japaneses document.
It lacked the description about joni direction enforcement so added
it. Addition to it, added html language specification and updated
target PostgreSQL versions and OS version.
Kyotaro Horiguchi [Tue, 16 Dec 2014 10:12:59 +0000 (19:12 +0900)]
Fix regression test according to the changes of PostgreSQL.
Row countings of EXPLAIN results seems to be changed. This follows the
change.
Kyotaro Horiguchi [Tue, 16 Dec 2014 09:45:22 +0000 (18:45 +0900)]
Removed a stale comment.
Kyotaro Horiguchi [Tue, 16 Dec 2014 05:45:13 +0000 (14:45 +0900)]
Fix to remove plpgsql query string on non-local exit.
Pl/pgsql query string stored internally squatted after an execution
error then pg_hint_plan afterward read it for hints and ignores really
given ones. This commit makes it erased at the end of transaction.
Kyotaro Horiguchi [Tue, 16 Dec 2014 04:20:02 +0000 (13:20 +0900)]
Fix about unexpectedly living plpgsql query string.
pg_hint_plan forgot to erase plpgsql query strings when it is NOT a
static statement, so it continued to read hints from the remenbered
wrong query string after using dynamic execution of a statement. This
commit makes it to be erased for any types of pl/pgsql statement.
Kyotaro Horiguchi [Tue, 16 Dec 2014 09:31:12 +0000 (18:31 +0900)]
Enhance debug output.
Added new debug output levels detailed and verbose. These let you see
more verbose information about what pg_hint_plan does internally.
Kyotaro Horiguchi [Wed, 17 Dec 2014 01:09:32 +0000 (10:09 +0900)]
Revert "Add new join method hint "NestLoop_NM" which means force nested loop"
This reverts commit
7f6e3ce1326751c36676653ef1e81893296860b6.
Kyotaro Horiguchi [Thu, 18 Sep 2014 05:23:15 +0000 (14:23 +0900)]
Add new join method hint "NestLoop_NM" which means force nested loop
but inhibit materialize of the inner relation.
Kyotaro Horiguchi [Wed, 17 Sep 2014 08:13:57 +0000 (17:13 +0900)]
Fix a bug that index restriction by name doesn't work for UPDATEs on
inheritance parent.
Inheritance planner doesn't claim for the inheritance parent relation,
so pg_hint_plan_get_relation_info cannot have the chance to prepare
the index spec information for chlid relations. This change make the
function to try to find parent relation even if it is called for the
children with inhparent == false from the first.
Kyotaro Horiguchi [Wed, 17 Sep 2014 08:10:18 +0000 (17:10 +0900)]
Remove unused struct member.
Kyotaro Horiguchi [Fri, 5 Sep 2014 04:35:55 +0000 (13:35 +0900)]
RPM automation
Kyotaro Horiguchi [Thu, 4 Sep 2014 09:04:56 +0000 (18:04 +0900)]
Change verision to 1.1.1. Change target version to 9.4.
Kyotaro Horiguchi [Thu, 4 Sep 2014 08:55:38 +0000 (17:55 +0900)]
Make codes copied from PostgreSQL catch up to 9.4
Kyotaro Horiguchi [Thu, 4 Sep 2014 02:03:08 +0000 (11:03 +0900)]
Interim support for 9.4beta2. pg_stat_statements.c is replaced with the latest version. Regtests are rearranged and stabilized.
Kyotaro Horiguchi [Fri, 5 Sep 2014 03:34:28 +0000 (12:34 +0900)]
Merge branch 'master' into PG93
Takashi Suzuki [Fri, 24 Jan 2014 04:07:00 +0000 (13:07 +0900)]
リグレッションテストを出力メッセージの変更に対応させた。
Takashi Suzuki [Fri, 24 Jan 2014 02:18:04 +0000 (11:18 +0900)]
hint_reportの出力メッセージを大文字始まりで統一した。
Takashi Suzuki [Tue, 21 Jan 2014 02:43:05 +0000 (11:43 +0900)]
コメントやメッセージを修正した。
Takashi Suzuki [Fri, 17 Jan 2014 02:49:38 +0000 (11:49 +0900)]
コピーライトを更新した
Takashi Suzuki [Fri, 17 Jan 2014 02:29:44 +0000 (11:29 +0900)]
Rowsで外部ファイルを扱う試験を環境に依存しないようにした。
外部ファイルのアドレスを特定の文字列に置き換えることで、環境に依存
しないようにした。
Takashi Suzuki [Fri, 17 Jan 2014 01:59:42 +0000 (10:59 +0900)]
Rows関連試験の予測結果を、全試験を流した場合のものに変更した。
Takashi Suzuki [Fri, 17 Jan 2014 01:52:44 +0000 (10:52 +0900)]
Rowsヒント句のメッセージ出力(R-3-6-*)の試験を追加した。
Takashi Suzuki [Fri, 17 Jan 2014 01:38:30 +0000 (10:38 +0900)]
Rowsヒント句の同じテーブル組合せ(R-3-5-*)の試験を追加した。
Takashi Suzuki [Fri, 17 Jan 2014 01:27:22 +0000 (10:27 +0900)]
Rowsヒント句の継承テーブル関連(R-3-4-*)の試験を追加した。
Takashi Suzuki [Fri, 17 Jan 2014 01:15:56 +0000 (10:15 +0900)]
Rowsヒント句の比率指定(R-3-3-*)の試験を追加した。
Takashi Suzuki [Thu, 16 Jan 2014 08:12:42 +0000 (17:12 +0900)]
Rowsヒント句の増減値指定(R-3-2-*)の試験を追加した。
Takashi Suzuki [Thu, 16 Jan 2014 07:44:21 +0000 (16:44 +0900)]
Rowsヒント句の絶対値指定(R-3-1-*)の試験を追加した。
Takashi Suzuki [Thu, 16 Jan 2014 07:43:38 +0000 (16:43 +0900)]
bug fix #3
Takashi Suzuki [Thu, 16 Jan 2014 06:57:30 +0000 (15:57 +0900)]
R-2-4-*の試験とR-2-5-*の試験を追加した。
Takashi Suzuki [Thu, 16 Jan 2014 06:52:16 +0000 (15:52 +0900)]
Rowsヒント句のRULEまたはVIEWの試験(R-2-3-*)を追加した。
Takashi Suzuki [Thu, 16 Jan 2014 05:08:55 +0000 (14:08 +0900)]
バグ修正にそって試験および予想結果を修正した。
Takashi Suzuki [Thu, 16 Jan 2014 05:02:23 +0000 (14:02 +0900)]
bug fix #2
件数補正のときに、全ての結合パスにおいて部分集合となってしまい、他
のRowsヒント句適用後に見積もり件数を再計算してしまった。。
理由
Rowsヒント句に存在しないテーブル組合せを指定すると、pg_hint_plan
内部で空のテーブル集合として扱っていたため。
Takashi Suzuki [Wed, 15 Jan 2014 09:41:04 +0000 (18:41 +0900)]
Rowsヒント句の件数の書式に関する記述を事実から要請に変更した。
Takashi Suzuki [Wed, 15 Jan 2014 08:43:52 +0000 (17:43 +0900)]
試験R-1-6-7の予測結果を正しい形に修正した。
Takashi Suzuki [Wed, 15 Jan 2014 08:42:11 +0000 (17:42 +0900)]
1ブロックあたりのテーブル数(R-2-2-*)の試験を追加した。
ただし、不合格
Takashi Suzuki [Wed, 15 Jan 2014 08:21:18 +0000 (17:21 +0900)]
複数の問合せブロックに関する試験(R-2-1-*)を追加した。
ただし、不合格
Takashi Suzuki [Wed, 15 Jan 2014 08:11:02 +0000 (17:11 +0900)]
core.cの各関数をPG93向けに更新した。
Takashi Suzuki [Wed, 15 Jan 2014 06:47:35 +0000 (15:47 +0900)]
マニュアルにRowsヒント句の仕様の一部と具体例を挿入した。
記入内容
・Rowsヒント句の補正の件数や倍率をstrtodで補正していること。
理由
・16進数でも指定可能なため。
Takashi Suzuki [Tue, 14 Jan 2014 02:02:30 +0000 (11:02 +0900)]
Rowsヒント句の引数の条件(R-1-*)の試験を追加した。
Takashi Suzuki [Thu, 9 Jan 2014 04:58:37 +0000 (13:58 +0900)]
機能試験表との試験番号のずれをなおした。
Takashi Suzuki [Tue, 24 Dec 2013 07:17:04 +0000 (16:17 +0900)]
スキャン方式試験の予測結果がSQLの変更に沿っていなかったため修正した
Takashi Suzuki [Tue, 24 Dec 2013 07:05:03 +0000 (16:05 +0900)]
スキャン方式試験のコメントが一部不適切だったため修正した。
Takashi Suzuki [Tue, 24 Dec 2013 07:00:56 +0000 (16:00 +0900)]
スキャン方式試験の番号のずれを修正した。
Shigeru HANADA [Tue, 24 Dec 2013 05:02:40 +0000 (14:02 +0900)]
Fix test numbers.
Shigeru HANADA [Mon, 9 Dec 2013 08:23:49 +0000 (17:23 +0900)]
PG9.3でのコスト値の変動を反映。
Shigeru HANADA [Mon, 9 Dec 2013 07:43:26 +0000 (16:43 +0900)]
PG92ブランチで追加したコメントに追従。
Shigeru HANADA [Mon, 9 Dec 2013 07:41:34 +0000 (16:41 +0900)]
Rowsヒントの追加に伴いリグレッションテスト項目の予想結果を追加。
9.2版からのコピーなので、細かいコスト値は正確ではない。
Shigeru HANADA [Mon, 9 Dec 2013 07:35:02 +0000 (16:35 +0900)]
PG92との差分から生成したパッチを適用。
ただし、PG9.3対応として加えた修正はそのまま残した。
Shigeru HANADA [Mon, 9 Dec 2013 05:19:48 +0000 (14:19 +0900)]
Add *.o to ignore file list.
Takashi Suzuki [Tue, 12 Nov 2013 07:43:05 +0000 (16:43 +0900)]
specファイル中の依存ファイル名を修正した。
RPM版インストール時の依存ファイル名に誤りがあったため
Takashi Suzuki [Thu, 7 Nov 2013 05:00:40 +0000 (14:00 +0900)]
Copyrightの未更新部分を更新した。
Takashi Suzuki [Thu, 31 Oct 2013 08:44:34 +0000 (17:44 +0900)]
コピーライトと一部の流用関数に更新漏れがあったため、修正した。
Takashi Suzuki [Thu, 31 Oct 2013 07:35:33 +0000 (16:35 +0900)]
スキャン方式の試験番号やタイトルが不適切だったため修正した。
Takashi Suzuki [Fri, 18 Oct 2013 06:50:59 +0000 (15:50 +0900)]
元のクエリの実行計画の変化に伴い、結合順序の試験の予測結果を修正した。
Takashi Suzuki [Fri, 18 Oct 2013 01:45:56 +0000 (10:45 +0900)]
元のクエリの実行計画の変化に伴い、結合方式の試験の予測結果を修正した。
Takashi Suzuki [Fri, 18 Oct 2013 01:24:00 +0000 (10:24 +0900)]
pg_hint_plan動作確認試験の予測結果をPG93の動作に合わせて変更した。
Takashi Suzuki [Thu, 17 Oct 2013 06:41:52 +0000 (15:41 +0900)]
スキャン方式の試験番号のずれを修正した。
Takashi Suzuki [Thu, 17 Oct 2013 06:36:21 +0000 (15:36 +0900)]
元のクエリの実行計画の変化に伴い、スキャン方式の試験の予測結果を修正した。