Thanks Dawa

It seems I was missing the -1 seed from RAND(), as I've been admonished for not reading Help.

I'll take a closer look at your suggestion tommorrow. I don't know HOW random I NEED to be; the user isn't exactly producing copious numbers of these per day. It was just an expediant to save his naming each s/s o/p. But worth considering.

Cheers

Terry

...
>
>Terry,
>
>I ran a test and it seems that the randomness becomes more apparent when you 'zoom' the result of rand() with larger number. The test prg generates 1000 random numbers with the specified zoom level into a cursor "ccc", then extracts the unique numbers from "ccc" and creates a new cursor "ddd", by comparing the recccount() of the two cursors you're able to tell the generated numbers are unique or not. The prg is self-explanatory:
>
...