Ohh, then locks wouldn't be released till transaction ends. If it's a must try to divide transaction to smaller pieces (a block does its job..end transaction, go next record).
Cetin
>I'm using Optimistic Table Buffering. After all the records are modified, the actual TABLEUPDATE() is wrapped in a TRANSACTION. Is this a problem?
>
>
>>If It's wrapped in a transaction you are implicitly locking each record. And If you're using novell it defaults to max 2000 locks, Net admin can change this but may not want too.
>>take it out of the transaction and release the locks
>>Jeff Hart
>>
>>>>>I have some code that SCANs a large number of records in a table and does some stuff with each record, including a REPLACE. On particularly large tables where a lot of records are processed, the server is giving a message that the record lock threshhold has been exceeded and the computer freezes up. Shouldn't the REPLACE command lock and unlock the records automatically? Why would a series of REPLACES cause a problem?
>>>>>
>>>>>Thanks.
>>>>
>>>>
Is it Novell ?
>>>>Cetin
>>>
>>>Yes, why?
