On Friday 17 October 2008 Scott Ribe's cat, walking on the keyboard, wrote:
> Older versions of PG were inefficient with larger numbers of elements in an
> IN query, and in fact would error out with something about lack of stack
> space if you used very many (a few hundred IIRC).
>
> 8.x something was supposed to have improved that. Using 8.3 recently, after
> an "oopsie" with some development data, I inadvertently confirmed that it
> works and performance is not too bad with >34,000 items in an IN clause ;-)
Interesting, since my queries sometimes expand to a few thousands of ORs (or
INs), and I had in fact stack problems and I had to expand it to around 20
MB.....
Luca