"SHELTON,MICHAEL (Non-HP-Boise,ex1)" <michael_shelton@non.hp.com> writes:
> Now, in the case of a select, then update/insert it will try an "optimistic"
> (hope I'm getting my terminology correct here) lock on the table (which
> will allow others to "read" the data or also apply "optimistic" locks --
> again for SELECT purposes -- but won't allow them to "change" or "write" the
> data locked -- less obtrusive this way, helps speed up concurrent access to
> tables). Then when you go to update whatever row, it would get an
> "exclusive" lock meaning no one is allowed to even "read" the data let alone
> "write" it. Again, the granularity (row vs page vs table) of the lock
> depends on the PG implementation itself.
>
> Hopefully one of the authors of PG will respond and correct/clarify anything
> I've said here.
Actually you're completely and utterly wrong. ;) PG doesn't work that
way, though other databases do.
See:
http://www.us.postgresql.org/users-lounge/docs/7.1/postgres/mvcc.html
-Doug
--
Let us cross over the river, and rest under the shade of the trees.
--T. J. Jackson, 1863