Proc logisitic gives error when column not present in the model

Question

I am trying to run a proc logistic regression. I have a lot of predictors which start with 'ST' and 'RF'. So I wrote the model statement like this

proc logistic data=x outest=y;
   model binary_variable (event='1')= age sex RF:  ST: lackfit;
  output out=x p=fitted_prob predprob=individual ;
run;

I have fed this statement inside a macro and running this macro for multiple scenarios. I encountered few datasets where there was no column starting with ST, and the proc logistic threw an error, which goes like this

WARNING: No variables found beginning with 'ST' in data set
NOTE: The SAS System stopped processing this step because of errors.

If there a way to handle such exceptions? Some kind of parameter inside proc logistic? Thanks!

user667489 user667489 · Accepted Answer · 2016-12-06T11:54:41

Within your macro, add some logic that detects whether each input dataset contains any columns starting with these prefixes and add the corresponding wildcards to the model statement only if at least one such column is present. You can get this information from sashelp.vcolumn or the proc sql equivalent dictionary.columns.

Proc logisitic gives error when column not present in the model

1 Answers