topic Re: List of Most Recent Created Variables in SAS Procedures

List of Most Recent Created Variables

Altal — Sun, 02 Mar 2014 16:19:44 GMT

Hi All,

Does anyone know a way in proc content or SQL to get a list of the "most recent variables created in the last data step"?

I have a big data set with large number of variables, and I'm interested to retrieve names of variables created in the last data step.

Thanks for help

Re: List of Most Recent Created Variables

Tom — Sun, 02 Mar 2014 17:47:15 GMT

That doesn't really make any sense. When you create a data set you create ALL of the variables for that data set. Even if some of them are coming from the inputs. The order of the variables will be determined by the order that the compiler notices the variables, not whether the step has made any modification to the values.

To see the most recently created data set you can use the &SYSLAST macro variable.

proc contents data=&syslast varnum ;

run;

Re: List of Most Recent Created Variables

Altal — Mon, 03 Mar 2014 00:41:50 GMT

Thanks for your response. Correct. I'm using SET statement on a dataset that has a large number of variables to create a new dataset with even larger number of variables created using a macro based on certain naming criteria I'm implementing. I was interested in retrieving the variable names that were created the last.

I was using the naive approach below but was wondering if there is a smarter way...!

data X;

set Y;

..... macro to create large number of variables;

run;

proc sql number;

select name , label

from dictionary.columns

where lowcase(Libname)='work' and lowcase(memname)='X'

and name not in

(select name from dictionary.columns

where lowcase(Libname)='work' and lowcase(memname)='Y');

quit;

Re: List of Most Recent Created Variables

Quentin — Mon, 03 Mar 2014 03:28:21 GMT

I think your approach makes sense.

As the number of libraries/datasets gets bigs, the performance of disctionary.columns can get slow. If that happens, could be faster to just run PROC CONTENTS on both datasets to make OUT= datasets with the variable names, and then compare those.

I haven't tried it, but it's possible you could get PROC COMPARE to do the work for you. If you run:

PROC COMPARE base=x (obs=0) compare=y (obs=0);

run;

I wonder if it would still compare the metadata?

Re: List of Most Recent Created Variables

Tom — Mon, 03 Mar 2014 14:11:57 GMT

That works. Add the LISTALL option to get the variables to list.

data class;

set sashelp.class ;

teen = age > 12;

run;

proc compare data=class(obs=0) compare=sashelp.class(obs=0) listall;

run;

Re: List of Most Recent Created Variables

Altal — Mon, 03 Mar 2014 14:25:34 GMT

Thank you all for your responses!