Re: BUG #16986: reindex error on ltree index
От | Alexander Korotkov |
---|---|
Тема | Re: BUG #16986: reindex error on ltree index |
Дата | |
Msg-id | CAPpHfdug2N+BsmQyPuuBhPyHQYe6NLX=Ta8KTdhnxVj=2KJxNA@mail.gmail.com обсуждение исходный текст |
Ответ на | BUG #16986: reindex error on ltree index (PG Bug reporting form <noreply@postgresql.org>) |
Список | pgsql-bugs |
.Hi, Mauro! On Thu, Apr 29, 2021 at 1:33 PM PG Bug reporting form <noreply@postgresql.org> wrote: > reindexing a table with ltree index this error is thrown: > ERROR: index row size 7720 exceeds maximum 8152 for index > "ltree_test_session_event_path_idx" > Follow script to reproduce error: > --------------------------------------------------------------- > CREATE TABLE "m3_dm"."ltree_test" ( > "session_id" varchar COLLATE "pg_catalog"."default" NOT NULL, > "session_event_path" "public"."ltree", > CONSTRAINT "ltree_test_pkey" PRIMARY KEY ("session_id") > ); > CREATE INDEX "ltree_test_session_event_path_idx" ON "m3_dm"."ltree_test" > USING gist ( > "session_event_path" "public"."gist_ltree_ops" > ); > insert into m3_dm.ltree_test > values ( > 'aaa', > text2ltree(repeat('abcdefghi.',480)||'abcdefghi') > ) > ; > reindex table m3_dm.ltree_test; Thank you for reporting. Generally, GiST isn't good at controlling maximum tuple length. If your tuple is long error can be raised randomly here and there. Your particular case is related to GiST fillfactor. Normal insertion doesn't take fillfactor into account but index build does. You can resolve this case by specifying fillfactor = 100. CREATE INDEX "ltree_test_session_event_path_idx" ON "m3_dm"."ltree_test" USING gist ( "session_event_path" "public"."gist_ltree_ops" ) WITH (fillfactor = 100); Besides fact that GiST need a general rework at controlling maximum tuple length, the error message looks really dumb. 7720 doesn't exceed 8152. I'm going to fix this. ------ Regards, Alexander Korotkov
В списке pgsql-bugs по дате отправления: