Richard Guo <guofenglinux@gmail.com> writes:
> The proposed patch can fix this error. However, I'm wondering if we
> could address it from the unique-ification side instead. If a Var
> we're trying to unique-ify is known to be equal to a constant, then we
> shouldn't need to unique-ify that Var
Yeah. I think this is an oversight in create_unique_paths(): it's
building an ORDER BY list without consideration for the possibility
that some of the entries are known to be constant. In fact, because
make_pathkeys_for_sortclauses will get rid of redundancies, this
example actually ends up with a unique_rel whose unique_pathkeys
are shorter than the unique_groupclause, which is pretty bogus.
Not sure about a good way to make it account for that though.
Detection of redundancies of this sort is kind of buried in
pathkey construction, but in this context we'd like to find out
earlier: we need to avoid attaching a new tlist entry if the
expression is constant, else we'll still get the failure.
regards, tom lane
