You can also eliminate that extra step and use curDups as derived table instead, e.g. put that select with () in the second select statement to use it as derived table.

Or you can keep the current approach in 2 steps.

>

>select min(mRec) as MinVs, MDate, COUNT(*) as cntDups ;
>from ch5 ;
>GROUP BY MDate ;
>HAVING COUNT(*) > 1 into cursor curDups NOFILTER
>SELECT T.*, IIF(T.mRec = curDups.MinVs, 1, 0) as KeepInfo, curDups.cntDups, curDups.minVs ;
>from ch5 T INNER JOIN curDups ON T.MDate = curDups.MDate into cursor curFinal nofilter
>

>
>>>Here it is. Using min(Vs), keepInfo found the first in alphabetical order. Using min(mRec), keepInfo found the top record in each group.
>>>
>>>

>>>mDate-----------Vs-----mRec-----keepInfo----cntdups---minVs
>>>05/28/2013-----menu------2---------1----------3---------2
>>>05/28/2013-----conso-----3---------0----------3---------2
>>>05/28/2013-----dire------4---------0----------3---------2  
>>>08/28/2013-----henr-----18---------1----------2---------2
>>>08/28/2013-----tran-----19---------0----------2---------2
>>>09/16/2013-----texa-----25---------1----------2---------2
>>>09/16/2013-----key------26---------0----------2---------2
>>>

>>>
>>>Thank you very much again!!!
>>>STeve
>>>
>>
>>You're welcome.