>1) Michael is looking for () is a the number of records with unique values in field1. Your query above returns the number of times each value appears.

>2) and in general DISTINCT slows performance.

>Now what will work in one select is:
>SELECT COUNT(DISTINCT field1) FROM delme
>Per the hackers guide.

If a rep works on 4 projects, they're still only 1 rep.

So should distinct be used if it's slower? Is it significantly slower than a two-step process for large tables? (my tables are likely to be <100 records anyway)

per hackers guide =?= not supported tomorrow, or just not documented anywhere else? (versions aren't an issue either, this company will probably never upgrade either)