On 02 Nov 2004 18:27:52 +0100, Harald Fuchs <[EMAIL PROTECTED]> wrote: > In article <[EMAIL PROTECTED]>, > > > >> Maybe you should explain what you're trying to achieve, not what you > >> think how to do it. > > > Sure. I am using MySQL to store packets for transmission. I want to > > send the oldest, highest priority packets first. Once I send it, I > > don't want to resend it for at least 15 seconds so I have time to wait > > for an acknowlegement. Based on external events, I may exit the loop > > before processing all the rows. In this case, I only want to update > > the timestamp on rows that I processed. Any rows that I haven't > > processed yet should keep the old timestamp. > > I see. Me thinks it's fundamentally unclean to iterate through a > result set and changing rows while doing so. You could remember the > pkeys of the rows processed and then do a single > > UPDATE ... WHERE pkey IN (...) > > after the loop, but maybe this might not be necessary at all: how > about SELECTing only a single row (with LIMIT), processing that row, > UPDATEing it, and all that within some loop? You're doing more > SELECTs that way, but with the proper indexes this might not be a > problem.
I was originally using "limit 1" to do this processing. However, I could get all the records in the same amount of time as it took to get one. So, for example, I can process thousands of records in four seconds or just one. I did resolve my problem, however. I created a temporary database connection to do the updates. ...Stephen -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
