On Fri, 9 Jul 2004, Peter Westergaard wrote: > > One idea that occurs to me, and it's a tradeoff from Torsten's idea (which > is to read the whole database, and parse out the unique rows), is to first > execute your "SELECT DISTINCT distinct_col FROM table", and then walk > through that, and for each one issue a "SELECT * FROM table WHERE > distinct_col = {value} LIMIT 1" (For some reason I can't remember the > 'LIMIT' syntax right now, so it might be LIMIT 0,1). More transactions with > the database, but less data actually retrieved from the database. And if > it's indexed by distinct_col, you should be fine. It´s not to start a war, but after having to hear so much about how good MySQL is. What you need is a database with sub-select capability, or to have a feature like: SELECT DISTINT ON(col1), col2, col5 FROM table1 WHERE... People say that they don´t need sub-selects (transactions, procedural languages, etc.) just because the don´t have an idea of the power that the hold. Just a par of cents on my part. -- 19:55:02 up 8 days, 11:32, 2 users, load average: 0.12, 0.15, 0.10 ----------------------------------------------------------------- Martín Marqués | select 'mmarques' || '@' || 'unl.edu.ar' Centro de Telematica | DBA, Programador, Administrador Universidad Nacional del Litoral ----------------------------------------------------------------- -- PHP Database Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php