Group Bys don't work in SQL like they do in VFP. You have to list every field in the SELECT list in the Group By clause.

I have a solution that uses correlated subqueries that is working very well.

Thanks for you time anyway.

>That's different story.

SELECT MIN(PK) AS PK, Value
>  FROM mytable
>  GROUP BY value
>-- OR
>SELECT MAX(PK) AS PK, Value
>  FROM mytable
>  GROUP BY value
>

>
>>I really don't care if it is the first or last. I really only want one.
>>
>>>Hi Dan,
>>>
>>>You have to define the meaning of "the first PK value" in terms of data in the table. IOW, which field defines the PK order and how.
>>>
>>>>I have a table that is using UNIQUEIDENTIFIER pk values. There is another character field that has duplicate values in it.
>>>>
>>>>For example (using integer pks for simplicity)
>>>>
>>>>

>>>>PK   Value
>>>>1    a
>>>>2    b
>>>>3    a
>>>>4    c
>>>>5    a
>>>>

>>>>
>>>>I want a query that will give me the first PK value for each unique "Value". So I will get back:
>>>>
>>>>

>>>>PK
>>>>1
>>>>2
>>>>4
>>>>

>>>>
>>>>How can I do this?