On Thu, 2009-03-19 at 08:41 +1000, Josh Marshall wrote:
> Maybe the INSERT INTO here can be a REPLACE INTO to avoid duplicates,
> otherwise ignore any duplicate errors as obviously that particular
> header already exists.

^^^ Or do a SELECT just prior to INSERT so that an error doesn't have to
be processed. Since this is only done during the maintenance time it
shouldn't impact speed too much

_______________________________________________
DBmail mailing list
[email protected]
http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Reply via email to