>Greetings!
>
>My question is not a *problem* with coding as much as a question about why one way would work faster than another.
>
>I have a routine that scans a rushmore-enabled table while searching other tables for a matching record which would result in returning a particular quantity. While experimenting, I found that two not-so-different methods can result in quite a bit of difference in run time.
>
>Here's the fast method:
>
>lnQty = -1
>SCAN WHILE lnQty = -1
> lnQty = GetQty2()
>ENDSCAN
>
>The GetQty2() routine returns (-1) if no match (no returned quantity) could be found.
>
>Now, here's a slight change that adds several minutes to the program's overall run time:
>
>SCAN
> lnQty = GetQty2()
> IF lnQty >= 0
> EXIT
> ENDIF
>ENDSCAN
>
>I suppose that the SCAN WHILE version takes more advantage of rushmore than the IF/ENDIF version, but is this what is really happening, or is it something else about how one command can work faster than another?
>
>After running across this, I now wonder how much of my other code may be in need of slight optimization.
>
>For example, elsewhere in this same program, I found a way to replace a LOCATE/CONTINUE loop with an embedded call to a SEEK inside a SCAN. The result is that instead of taking over TWO HOURS to complete, the program now runs in just over THREE MINUTES!
>
>Any input about identifying the differences of speed in relation to syntax would be very appreciated.
>
>Michael Reynolds
>Santa Barbara, CA


How about the very simple answer that #1 has fewer lines of code to execute.

DO WHILE !EOF()
* Processing
SKIP
ENDDO

is slower than

SCAN
* Processing
ENDSCAN

fewer lines of code and more processing being done by one command.