topic Re: Remove variables in a data set that does not contain a pattern and rename variables in New SAS User

Remove variables in a data set that does not contain a pattern and rename variables

NewUsrStat — Sat, 19 Aug 2023 15:28:44 GMT

Is there a way to remove columns that does not end with "_22" from a data set? Moreover, is there a way to remove "_22" suffix from column names from another data set?

Thank you in advance

Re: Remove variables in a data set that does not contain a pattern and rename variables

Kurt_Bremser — Sat, 19 Aug 2023 16:52:07 GMT

In PROC SQL, you can retriev variable names from DICTIONARY.COLUMNS. With SELECT INTO, you can create macro variables which contain the necessary lists for DROP and RENAME statements used in PROC DATASETS.

Re: Remove variables in a data set that does not contain a pattern and rename variables

NewUsrStat — Sat, 19 Aug 2023 18:41:49 GMT

Thank you kurt_Bremser. Is there an example or something like that to inspire to?

Re: Remove variables in a data set that does not contain a pattern and rename variables

Astounding — Sat, 19 Aug 2023 19:44:36 GMT

What ways do you know (if any) to get the names of variables in your data set?

Re: Remove variables in a data set that does not contain a pattern and rename variables

Kurt_Bremser — Sat, 19 Aug 2023 20:25:53 GMT

proc sql noprint;
select catx("=",name,substr(name,1,length(name)-3)), substr(name,1,length(name)-3)
into :renamestr separated by " ", dropstr separated by " "
from dictionary.columns
where libname = "WORK" and memname = "HAVE" and substr(name,length(name)-2) = "_22";
quit;

data want;
set have;
drop &dropstr.;
rename &renamestr.;
run;

Since you want to remove variables, PROC DATASETS can't do it, a DATA step is needed. My bad.

Re: Remove variables in a data set that does not contain a pattern and rename variables

ballardw — Sat, 19 Aug 2023 23:03:38 GMT

In my occasionally humble opinion, the question is why to do you have variables with names you don't want in the first place. Especially something ending in _22. That almost implies that you have 21 other names that are acceptable (ending in _1 _2 _3 ... _21).

When you have data set names you don't like and want different names then Proc Datasets will let you rename the variables in place.

Proc datasets library=Somelib;
   modify adataset;
      rename var_22 = var
             thatvar_22 = thatvar
       ;
run;
quit;

The example code will modify the dataset named Adataset in the library named Somelib by renaming tow variables.

Note this procedure uses a QUIT statement to end the procedure.