On Tue, 8 Dec 2009, niraj patel wrote:
> Thanks very much for the analysis. It does takes 17 sec to execute when
> data is not in cache.
It sounds like the table is already very much ordered by the workspaceid,
otherwise this would have taken much longer.
> What I would like to ask can partitioning around workspaceid would help?
> Or any sort of selective index would help me.
Depends on how many distinct values of workspaceid there are. I would
suggest that given how well ordered your table is, and if you aren't doing
too many writes, then there would be little benefit, and much hassle.
Now, you would have thought these coefficients would be integers, given that
we're working out integer results. Using a fraction would seem really
stupid. Well, I'm quite willing to be stupid here - in fact, I'm going to
use complex numbers. -- Computer Science Lecturer
In response to
pgsql-performance by date
|Next:||From: Robert Haas||Date: 2009-12-08 15:02:05|
|Subject: Re: performance penalty between Postgresql 8.3.8 and 8.4.1|
|Previous:||From: niraj patel||Date: 2009-12-08 14:27:35|
|Subject: Re: Optimizing Bitmap Heap Scan.|