>>Hi everybody,
>>
>>How can I randomly select n (110) codes from some huge table, so these codes would be unique? I'm trying to create a test case now...
>
>

>*   Assumes V. 7 for RW cursor, but can emulate in prev versions.
>
>SELECT cCodes, .F. AS lSelected ;
>   FROM BigTable ;
>   INTO CURSOR UniqueCodes readwrite ;
>   GROUP BY cCodes
>
>lnCodes = _Tally
>
>SELECT UniqueCodes
>
>*  This approach to using RAND() will give randomized numbers, but
>*  the same sequence each time it is run (this is helpful for testing
>*  to determine that changes in execution speed are due to algorithm
>*  changes rather than the distribution of the values).
>
>*  If you want a different series between runs, you can issue a
>*  preliminary call to RAND (-1) which will seed with a value
>*  from the system time clock. For that, uncomment the line below
>
>*  = RAND (-1)
>
>lnSelected = 0
>DO WHILE lnSelected < 110
>   GO lnCodes * RAND () + 1
>   IF NOT lSelected
>       REPLACE lSelected WITH .T.
>       lnSelected = lnSelected + 1
>   ENDIF
>ENDFOR
>
>SELECT cCodes ;
>   FROM UniqueCodes ;
>   WHERE lSelected ;
>   INTO CURSOR SelectedCodes
>
>

I'd rather create cursor SelectedCodes and insert records into it from within the Do While lnSelected<110 loop, so you wouldn't have to run a select against a long table in the end. But other than that, this seems to be the ticket.