>>>> Hello all,
>>>>
[quoted text clipped - 33 lines]
>>
>> Br@dley
> Two potential problems here are that with 500,000 names (as the OP
> says) this is not going to be fast - I could imagine it would pretty
> much grind to a halt.
Sure. There are lots of ways to skin a cat. The most elegant/complex of SQL
statements may just be too slow realistically.
> Secondly, if you are ordering the people
> alphabetically and selecting in any kind of structured way, then the
> results cannot be said to be random.
Ah, I missed the random part and only saw the second part (and the subject)
where he stated he wanted every 25th record...
> Now I'm not a mathematician, and don't want to enter a huge debate on
> what is random, but wouldn't this be better?
> SELECT TOP 20000 * FROM People ORDER BY Rnd([ID])
Yep, that'd be the easiest way to get a random sample.
(I like doing things dynamically so I'd probably pass a parameter such as
the percentage or number of records I wanted returned).
Thanks for pointing that out.

Signature
regards,
Br@dley