>>I am converting my updates to my table from the legacy scatter / gather technique and trying out buffering. I'm a litle confused though when I issue tableupdate(). If I understand this technique correctly, when I have (OPT 3) set on a table either in code or in the bufferoverride property of a DE, when an edit is made it is made to the table directly, and the only time locking is performed is when the pointer is moved, or the tableupdate() command is issued.
>>
>>When the lock is placed on the record edited when tableupdate() is issued, does it automatically unlock that record so other users can edit that record after the tableupdate() is complete or do I have to issue an unlock after tableupdate() succeeds?
>>
>>Thanks.
>>
>>john.
>
>Also, will the buffering behave unusual if I set step on an step through the program one line at a time? When I did this, the tablerevert() wouldn't restore my the orignal value of field I edited, but if I ran the program straight through the tablerevert() worked fine.
>
>Thanks!

Hi John,

As I understand it, TABLEUPDATE() will lock the necessary record(s), update the table, then release the lock(s). TABLEUPDATE() should return a .T. or .F. as to whether the update of the record(s) was successful. If it failed you can issue AERROR() to determine what went wrong. BufferModeOveride in the Help gives some info along with TABLEUPDATE().

As to the second question, I don't think setting step on should affect the operations. However, i have noted the sometimes things don't happen exactly the same when the debugger is running. This has been something that has bothered me. :-)

Bill