Steven Grimm <sgr...@thesegovia.com> writes:
> It is puzzling that if, as suggested by someone else in the thread, I
> expand IN(a,b,c) to (x = a OR x = b OR x = c) it gets substantially
> faster, though still obviously falls afoul of the problem you describe
> above (~4 seconds instead of ~6 seconds). Should those two be equivalent?
The parser actually will do that expansion for you, when the IN-list items
contain variables ... but its definition of "variable" for this purpose is
"contain_vars_of_level(expr, 0)" so the outer-level Vars you've got in
this LATERAL subquery formulation don't trigger that behavior. I seem to
remember writing it that way intentionally, but this example makes me
think maybe excluding outer-level Vars wasn't such a hot idea. It will
remain a ScalarArrayOpExpr even if the query later gets flattened to the
point where the Vars aren't outer-level anymore, which is probably not
what we want it to be.
regards, tom lane
--
Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-general
