]> granicus.if.org Git - postgresql/commit
projects / postgresql / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 421467c) | patch
Further thought shows that has_distinct_on_clause() needs to take much
authorTom Lane <tgl@sss.pgh.pa.us>
Tue, 31 Jul 2001 20:16:33 +0000 (20:16 +0000)
committerTom Lane <tgl@sss.pgh.pa.us>
Tue, 31 Jul 2001 20:16:33 +0000 (20:16 +0000)
commit0889bd00bd365ca252d75118378171fe5fd48289
treed027d973d075674f6c1428c5bbc147e81ea9aab8tree | snapshot
parent421467cdc8231a9da7ae6116ce6030d7d756f137commit | diff
Further thought shows that has_distinct_on_clause() needs to take much
more care with resjunk tlist entries than it was doing.  The original
coding ignored resjunk entries entirely, but a resjunk entry that is
in either the distinctClause or sortClause lists indicates that DISTINCT
ON was used.  It's important for ruleutils.c to get this right, else we
may dump views using DISTINCT ON incorrectly.
src/backend/optimizer/util/clauses.c diff | blob | history
Unnamed repository; edit this file 'description' to name the repository.