>>select distinct rep, .t. as SA_flag from myTable ; where pub = "SA" ;union ;
>select distinct rep, .f. as SA_flag from myTable ;
> where rep not in (select distinct rep from myTable where pub="SA") ;order by 1
>
>Assuming a significant size "myTable", how would that statement compare with:
>Select distinct rep, .t. as SA_Flag from myTable where pub="SA";
>union;
>Select distinct rep, .f. as SA_Flag from myTable where NOT (pub="SA")
>
>It seems the first statement would be querying the same set twice.
>
>I thought there was an "in-line" way to get a unique value from a group, but i guess the two-step union will work (almost) as effectively.

---

Michael,
I thought it but I think it won't work. Why, because there could be records with pub # "SA" with same rep those having pub="SA". Union wouldn't handle it becuse of the unmatching SA_Flag.

rep   pub
---   ---
1     
1     SA
1
2
3
4     SA
4
5
5     SA

*would produce
rep   SA_Flag
---   ---
1     .T.
1     .F.
2     .F.
3     .F.
4     .T.
4     .F.
5     .T.
5     .F.

For performance and one pass there could be other clever SQLs :) Like :

Select rep, iif(sum(iif(pub="SA",1,0))>0,.t.,.f.) as SA_Flag ;
 from myTable ;
 group by 1

Cetin