topic Remove observations from one dataset based upon a list in another in SAS Procedures

Remove observations from one dataset based upon a list in another

yeaforme — Sun, 27 Oct 2013 14:44:13 GMT

I have two datasets. Dataset A has all the raw data where each subject ID code has multiple (sometimes thousands) of observations for it. Dataset B has a list of subject ID codes that need to be removed from Dataset A (i.e., if the ID code is in Dataset B then all observations linked to the ID code need to be removed from Dataset A).

I'm thinking it would seem like a job for proc sql, but don't know how to go about it. Thanks!

Re: Remove observations from one dataset based upon a list in another

Jagadishkatam — Sun, 27 Oct 2013 14:53:34 GMT

proc sort data=dataset_A;

by id;

run;

proc sort data=dataset_B;

by id;

run;

data want;

merge dataset_A(in=a) dataset_B(in=b);

by id;

if a and not b;

run;

by proc sql;

proc sql;

create table want as select a.* from dataset_A as a, dataset_b as b where a.id^=b.id;

quit;

Thanks,

Jagadish

Re: Remove observations from one dataset based upon a list in another

yeaforme — Mon, 28 Oct 2013 04:02:24 GMT

The SQL code creates a cartesian product that takes forever to run before timing out after it has consumed all available memory, but the data merge code above it works well.

Thanks!