I'm trying to use proc surveyselect to conduct a random drawing. Entrants can have multiple entries, and I have a table that lists each entry. The problem is I'm getting duplicate winners because of the multiple entries. Is there a way to exclude duplicate observations by a certain column (Entrant_ID in this case)? Consider this data set, but in actuality I have 100 dept names, and about 3,000 entrant IDs (most of which are duplicates. There are really about 90 unique entrant IDs per Dept Name. Entrant ID's will not duplicate over different dept's). Also, I will have a different actual sampsize for each dept, but I understand there is a way to make that reference a separate table, so I will be looking to do that. Dept_Name Entrant_ID Dept 1 JRTHAL Dept 1 TLSMIT Dept 1 JRTHAL Dept 1 VLNEW Dept 2 MRREL Dept 2 MNJON Dept 2 MRREL Dept 2 MNJON Dept 2 NWCON Dept 3 JRMCC Dept 3 ADLON Dept 3 JRMCC Dept 3 ADLON Dept 3 BFPIT Dept 3 BFPIT Code I am using: proc surveyselect data=work.random_draw_entries out=all_winners method=srs reps=1 sampsize=10; strata dept_name; run; Also, I'm not tied to using surveyselect, and I have many ways to manipulate the data if you have suggestions. Thanks!
... View more