2006/12/4, Ian Harding <harding.ian@xxxxxxxxx>:
On 11/13/06, Scott Marlowe <smarlowe@xxxxxxxxxxxxxxxxx> wrote: > On Mon, 2006-11-13 at 15:36, novnov wrote: > > OK, thanks everyone, I gather from the responses that postgres performance > > won't be an issue for me then. If MS SQL Server and Postgres are in the same > > ballpark performance-wise, which seems to be the upshot of your comments, no > > problem. I'd only have worried if there was something like the major > > difference between the two with more complicated queries. I am puzzled by > > the commentor's post to the article, it could be FUD of course but didn't > > particularly sound like the commentor was anti pgsql. > > I will say this. Most other databases are more forgiving of bad > queries. Make a bad query and postgresql is more likely to punish you > for it. Amen. When I migrated from MSSQL to PostgreSQL (4 years ago), I found out exactly how seriously MS SQL coddles you when it comes to its "Oh, I know what you really meant" query planning. I committed some sins MS SQL covered up nicely and PostgreSQL flat out crawled when presented to it. However, I suspect that if I tried those bad queries with a current version of PostgreSQL they would run much better, given all the work that has been put in over the last few years.
I've seen the exact same behaviour last year with pg 8.1 vs. MS SQL 2k. It was an unexpected shock, but it's really not that hard to make pgsql run much faster. There are simply things which pgsql executes painfully slow if you don't write them the way the server expects you to. This hasn't changed in 8.1, but then again, it's not nearly the biggest problem I have with this specific RDBMS. ;) t.n.a.
