On Tue, Jun 29, 2004 at 01:37:30 -0700,
Chris Cheston <ccheston(at)gmail(dot)com> wrote:
> ok i just vacuumed it and it's taking slightly longer now to execute
> (only about 8 ms longer, to around 701 ms).
> Not using indexes for calllogs(from)... should I? The values for
> calllogs(from) are not unique (sorry if I'm misunderstanding your
If you are hoping for some other plan than a sequential scan through
all of the records you are going to need an index. You can have an
index on a column (or function) that isn't unique for all rows.
In response to
pgsql-performance by date
|Next:||From: Bruno Wolff III||Date: 2004-06-29 13:44:06|
|Subject: Re: no index-usage on aggregate-functions?|
|Previous:||From: Michal Táborský||Date: 2004-06-29 12:30:30|
|Subject: Slow INSERT|