Claudio,

I get the same result as you. Moreover, if I test with 10,000 records the result query is extended as expected: 17, 273, 529, 785, 1041, 1297, ..., 9745. This is very weird. I would report this to MS if I were you, either we're missing something obvious or it's a bug. BTW I am VFP6/SP3 on Win98SE.

-Rick

P.S. Move your "replace all" statement out of the loop, your test will run *lots* faster.


>Hi, everybody!
>
>I need your help... I'm with a problem that is ocurring also with some other friends (and with some other, not)...
>
>Do as follows:
>
>* Create a table called Test1 with the structure:
>Field: iID
>Type: Integer
>
>* Create a table called Test2 with the structure:
>Field: iID
>Type: Integer
>
>* Create any FOR/NEXT to fill these tables with some 1000 records
>For example:
>SELE TEST1
>FOR i=1 TO 1000
> appen blank
> repl all iID with recno()
>ENDFOR
>
>Do this in both tables.
>
>Well, these two table has the same structure and the same content (records and value)
>
>Now, run the following command:
>
>SELECT * FROM TEST1 WHERE IID NOT IN (SELECT IID FROM TEST2)
>
>We're trying to descover which records is in the Test1 and don't have a corresponding record in Test2. It must return Zero records, because they're exact the same. However, here comes the problem: in my SQL, returns a cursor with the folling records: 17, 253, 509 and 765. Do you see? It begun with 17, and goes adding 256... 17+256=253... 253+256=509.. etecera...
>
>After did a lot of attempts, I resolved to change the type of field IID of table Test2, to Numeric 4,0... well, it does work... I'm sure that it's a problem with Integer Fields.
>
>Could anybody explain this for me?
>
>Thanks in Advance
>
>[]'s
>
>Lassala