Michael,

In a FPD 2.6 app, I would judge this code to be reasonably fast!

What is it that is slow?. . . A *LOT* of indexes would slow it down, even more so *IF* there are several people executing the same code on fairly large numbers of records. But if this code is basically restricted in its usage, and if the number of TAGs is small, I'd venture you won't get much faster.

I'd also venture that there is *NO WAY* to avoid Locks of some kind.

good luck,
Jim N

>>In my ongoing saga of legacy code in FP2.6 for DOS (ugh), I found someone doing something like:
>>
>>Scan
>> Scatter Memvar
>> Insert Into Booked From Memvar
>>Endscan
>>
>>The reason I was given for doing this was that they could not allow the record/table to be locked as would be the case with the append command. This code is incredibly slow. Is there a way around this witout locking tables?
>
>Asking why the record/table couldn't be locked would be my first question. You aren't by any chnace working on a Banyan Vines LAN are you?
>
>In my FP2.x experience on Banyan, yes you could get into a sticky situation if during the append (table header is locked), the LAN connection went down, or some other catastrophic occurance (ungracefully exiting the program because user thought it was hung). If this happens then you've got a permanent lock on the table header, essentially making it a read-only file. The only way to recover is to reboot the server, very ugly. I did resort to the SCAN .. SCATTER for appends when the LAN was very wiggy. It was either that, or reboot the server just about every day (making me a VERY unpopular person, in the extreme!).
>
>HTH,
>Valerie