Hi Kevin

How much faster was this approach?

>Hey Mike,
>You are right, this is a much better way to do this. Sometimes you get moving forward and can't see the forest through the trees! Thanks for your help.
>Kevin
>
>>Hi Kevin
>>
>>If a piece of code is taking even 20 minutes, I'd be unhappy ;).
>>
>>Do you want to pull new data into your existing data? If so, you won't need to delete anything (implying a later process to get the undeleted records via pack or another SQL)
>>
>>How about flipping the entire process over?
>>
>>

>>SELECT * ;
>>  FROM NEW ;
>>  WHERE ;
>>    ssn + holder_num + DTOS(date_rem) NOT IN ;
>>      (SELECT DISTINCT ssn + holder_num + DTOS(date_rem) FROM old) ;
>>  INTO CURSOR ;
>>    c_KeepThese
>>