>>>>>>

>>>>>>CREATE CURSOR Test ( iInt I )
>>>>>>INSERT INTO Test ( iInt ) VALUES ( 1 )
>>>>>>INSERT INTO Test ( iInt ) VALUES ( 2 )
>>>>>>GO TOP IN Test
>>>>>>REPLACE iInt WITH 5 FOR iInt < 3 IN Test && 2 replacements
>>>>>>

>>>>>
>>>>>See message#1624599 :)
>>>>>
>>>>>BTW: GO TOP is superfuous. ALL and implicit ALL in FOR will do anyway.
>>>>
>>>>The purpose of the code was to demonstrate the default scope when the FOR clause is included. If it was NEXT 1 as the VFP help file says, then only the top row would get REPLACEd. Since both rows are affected, it proves that the default scope is ALL.
>>>
>>>But for the ones not knowing it might be even more surprising that it will run through all records from the last (second) record too.
>>>I think the problem comes from the SCAN ENDSCAN use of FOR WHILE and scope. In SCAN it is clear what FOR means. So instead of
>>>

>>>SCAN FOR expr
>>> REPLACE ...
>>>ENDSCAN
>>>

>>>the
>>>

>>> REPLACE ... FOR expr
>>>

>>>was introduced.
>>
>>Hmm, that's going way back in time. I know for sure the REPLACE command has been around since early dBASE days. SCAN was introduced in FPD if memory serves. ISTR the FOR clause of REPLACE has also been around forever, I believe it predates SCAN, so it wasn't introduced to fix anything to do with SCAN.
>
>I know this is from the days where the user was hacking the code in command line and a SCAN construct was inconvenient so yes, it might be there from start, and REPLACE the more basic command.
>Software archaeology. :)
>
>Please define: ISTR

I Seem To Recall.