DIS 4.4 SAS 9.3 New Developer So, I've got a dataset and I want to 'select distinct' only on a subset of variables. If the observation is distinct, then I want all of it's variables to pass through. For example, I have 30 variables, but I only want to check 5 of them and if those 5 are the same between two (or more) observations, than I only want one of the observations to go through, but the values for all 30 variables. I would guess I would use an Extract transformation with the option of select distinct set to 'Yes'. However, this isn't doing it. When I only map the 5 variables to the outgoing dataset, then it works, but I only have 5 variables per observation and not the 30 I desire. Did I make any sense? Basically: If I have Var_1, Var_2, Var_3, ...Var_30, but want to only select on Var_3, Var_5, Var_6, Var_10, and Var_22. I think doing a 'group by' could also do the trick, but also doesn't seem to: Select Var_1, Var_2, Var_3, ... Var_30 from TABLE group by Var_3, Var_5, Var_6, Var_10, Var_22 I still end up with the same number of observations. Also, is there a way to output a dataset (even a temporary one) of all observations that were not distinct so that I can view that group as well? Say if I have 200 observations, but after selecting distinct, I have 175...is there a node/transformation I can use to view those 25 that were 'rejected'? Thanks!
... View more
