Well, yes and no...

Firstly, because of VFPs cacheing, there may be no perceptible difference.
Secondly, when actually using buffering (row or table), there likely will be no perceptible difference.

It is true, though, that dBaseIII+ was certainly much faster when doing a single replace rather than 3 (in the case shown).

I'm betting no perceptible difference, in which case 3 separate replaces or a single replace using 4 (not 3) lines to do it both provide the same ease of code maintenance.


>It would be fairly simple to create a test program for this.
>
>
>>Thank you Craig,
>>
>>You confirm what I was thinking... My boss didn't believe me and I have no proof to give. For me it seems to be an evidence !
>>
>>
>>Thanks again