Kyotaro HORIGUCHI <> writes:
> Oops! I found a bug in this patch. The previous v8 patch missed
> the case that build_index_pathkeys() could build a partial
> pathkeys from the index tlist.

TBH I think that's barely the tip of the iceberg of cases where this
patch will get the wrong answer.  It looks like it thinks that *any*
Var belonging to the indexed relation must be the same one indexed
by the index.  Also, I don't see it doing anything to check the ordering
of multiple index columns --- for instance, an index on (a,b) and one
on (b,a) would both pass or fail identically, though surely only one
should match "ORDER BY a,b,...".  Also, what's with the success return
before the loop:

+       if (list_length(pathkeys) == list_length(root->query_pathkeys))
+               return true;

At this point you haven't proven *anything at all* about whether the
index columns have something to do with the query_pathkeys.

                        regards, tom lane

Sent via pgsql-hackers mailing list (
To make changes to your subscription:

Reply via email to