Re: [GENERAL] select statement against pg_stats returns

Christopher Kings-Lynne wrote:
> > Why?  You can reconstruct it with a simple "ANALYZE" command.  Dumping
> > and restoring would mean nailing down cross-version assumptions about
> > what it contains, which doesn't seem real forward-looking...
> 
> I seem to recall that people like that kind of thing so that the dump is 
> really the current state of the database.
> 
> Also, I believe big db's like DB2 and Oracle do such a thing.
> 
> I just recall it being discussed some time ago...

I have heard of dumping stats so you are sure your production db has the
same stats as your test database, but with ANALYZE so fast, and our
optimizer so good, I don't see a use case for us.

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
359-1001+  If your life is a hard drive,     |  13 Roberts Road +  Christ can be your backup.        |  Newtown Square,
Pennsylvania19073