Okay, I feel like I need to clarify some things. I do have a UNIQUE INDEX constraint on those columns, so the other user won't actually write the same value another time, but it will fail at a level which it should not.
I don't want to use AUTO_INCREMENT because it's not portable. My application should work on MySQL and SQLite (and maybe someday it will also run on many other systems - today, incompatibilities are just too big). Here's another example: SELECT COUNT(*) FROM table WHERE name = ? -- a short delay which is long enough for a concurrent request :( UPDATE table SET name = ? WHERE id = ? I do the first query to find out whether my new name is already assigned. Each name can only appear one time. If I just try and update the row, the query will fail, but I don't know why. All I could do is try and parse the error message, but this will by DBMS-dependent. I'd like to do it in a way so that I can tell the user whether the name was not unique or there was another error. But this case should be detected separately. I'll have a look at those isolation levels though. Maybe it's what I'm looking for. -- Yves Goergen "LonelyPixel" <[EMAIL PROTECTED]> Visit my web laboratory at http://beta.unclassified.de -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
