>>
>>Larry- I think what you are trying to do can be accomplished with one statement:
>>
>>SELECT count(*), distinct *;
>>FROM masterfile;
>>INTO CURSOR countfile;
>>WHERE inactive "Y";
>>GROUP BY agegrp, district
>>
>>This doesn't directly apply to your question, but could possibly reduce the time it takes to create your countfile table.
>
>But the index is agegrp+district. Grouping this way works in the same way as that index would?

The group by essentially says (in pseudo-english) "create new total record when agegrp AND district change", therefore the index WILL work. Besides, you can't group by agegrp+district....at least I've never gotten it to do it.