>>>>>If so, this combination should do it (note slight change in first SQL)

SELECT MIN(PrimaryKey) AS pk, SecondaryKey AS sk ;
>>>>>FROM Table INTO CURSOR TEMP ;
>>>>>GROUP BY 2
>>>>>* now choose smaller of two for first column
>>>>>SELECT EVAL("MIN(pk,sk)") AS pk, sk FROM temp

>>>>
>>>>Thanks to all who helped me on this.
>>>>
>>>>I chose this approach. I guess all approaches sent are good as well. And, the speed is good as well.
>>>
>>>As usual, the Fox gives us *many* ways to solve a problem.
>>
>>And here's one more. The non-SQL approach:
>>
>>INDEX ON SecondaryKey TO whatever UNIQUE
>>COPY FIELDS PrimaryKey,SecondaryKey TO whatever
>
>But that would not always give the *lowest* PrimaryKey for each SecondaryKey, which was part of the requirement.

I must have read it wrong. I thought it was the FIRST instance of PrimaryKey for each SecondayKey that was being sought, not the LOWEST instance.