Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
> What do you think of making vacuum a no-op on a table that has datminxid
> = FrozenXid?
I think this is probably OK for plain VACUUM and VACUUM FREEZE, but not
for VACUUM FULL, since a frozen table isn't necessarily fully packed.
Also, for VACUUM ANALYZE, the ANALYZE step is likely still needed?
regards, tom lane