On Tue, 2024-07-23 at 21:37 -0400, Robert Haas wrote:
> In my experience, sorting is, overwhelmingly, the problem.
I strongly agree.
> That we have versioning information that someone could hypothetically
> know how to do something useful with is not really useful, because
> nobody actually knows how to do it
Including me. I put significant effort into creating some views that
could help users identify potentially-affected indexes based on
collation changes, and I gave up. In theory it's just about impossible
(consider some UDF that constructs queries and EXECUTEs them -- what
collations does that depend on?). In practice, it's not much easier,
and you might as well just reindex everything having to do with text.
In contrast, if the problem is CTYPE-related, users are in a much
better position. It won't affect their primary keys or most indexes.
It's much more tractable to review your expression indexes and look for
problems (not ideal, but better). Also, as Peter points out, CTYPE
changes are typically more narrow, so there's a good chance that
there's no problem at all.
Regards,
Jeff Davis