My FAQ # 8129 is dealing with the exact same problem. Take a look.

>My program had a bug and when inserted duplicate records. What SQL can I use to delete all the duplicates, but one?
>
>To find the duplicates I would use the sql below, but only want where cnt_text > 1. Then I want to delete all the records but the lowest sequence number (seq).
>
>select key, seq, desc, comp_dt, count(text) as cnt_text from tickler ;
>where desc = "Letter" and comp_dt = date()-1 ;
>order by key, seq ;
>group by key
>
>thanks