(Apologies in advance for not indenting.  I need to noodle with css or something).

Say you import an xlsx file…

libname logit '\\psf\Home\Desktop\';
proc import datafile="\\psf\Home\Desktop\bc.xlsx" out=logit.bc dbms=xlsx replace;
run;

…and it has no header row.  SAS will come up with its own variable names and labels:

ods pdf file='\\psf\Home\Desktop\01.pdf' style=SASdocprinter;
options nodate;
title Proc Contents One-After Proc Import;
proc contents data=logit.bc;
run;
title;
ods pdf close;

Ick.

Use PROC Datasets to rename and relabel variables:

/* Renaming and labeling variables */

proc datasets library=logit;
modify bc;
rename _24=wife_age  _2=wife_edu  _3=hus_edu  _3_1=num_child  _1=wife_rel 
_1_1=wife_work   _2_1=hus_occ  _3_2=sol  _0=media  _1_2=bc;
label wife_age="Wife's Age"  wife_edu="Wife's Education"  hus_edu="Husband's Education"num_child='Number of Children'  wife_rel="Wife's Religion" 
wife_work="Wife Work Outside Home?"  hus_occ="Husband's Occupation"
sol='Standard of Living'  media='Media Exposure'  bc='Contraceptive Method';
run;
quit;

ods pdf file='\\psf\Home\Desktop\02.pdf' style=SASdocprinter;
options nodate;
title Proc Contents Two-After Proc Datasets;
proc contents data=logit.bc;
run;
title;
ods pdf close;

That’s better!

It’s best not to use a data step simply to change the name of a SAS dataset or columns in a SAS dataset. This requires the entire dataset to be rewritten. Insead use proc datasets, which changes the names without needing to read the observations in the dataset.

Thanks!

---

 

Original dataset from “Contraceptive Method Choice” courtesy of the UCI Machine Learning Repository.