On Wednesday 7. June 2006 00:10, Jim C. Nasby wrote: >Moving to -general. >From the bottom of that page: > > SELECT * FROM sources INTO src WHERE source_id = $1; > >SELECT * is generally something to avoid. You end up shoving around > more data than needed. Granted, in this case it's only getting shoved > down into plpgsql, but it's still extra work for the server. I know that. But the table is only four columns wide, and all the columns matter in this query. Eventually I'll remove such things as "SELECT * FROM ..." which really is a bad habit. >Also, the commentary about how MySQL is faster isn't very clear. Are > you using MySQL as some kind of result cache? When you get to running > actual concurrent access on the website, you could well find yourself > very disappointed with the performance of MyISAM and it's table-level > locking. There's probably also some gains to be had on the PostgreSQL > performance. I may have been a little unclear here. My production database is PostgreSQL, as it quite clearly is the better choice of the two, in particular wrt data integrity. My Web presentation software is quite a different matter. It's running at a web hotel that's only offering MySQL for a database. I find MySQL with MyISAM quite sufficient for that use, as its only purpose is to serve up simple selects quickly. The reason why the generation of eg. the family sheet is faster in the MySQL web context than in my production environment, is that I'm really comparing apples and potatoes here. The Web database has a much flatter and denormalized structure, due to the fact that there's no editing. The entire Web database is repopulated from scratch every time I do an update. -- Leif Biberg Kristensen | Registered Linux User #338009 http://solumslekt.org/ | Cruising with Gentoo/KDE
