- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Posted 08-06-2009 02:51 AM
(7620 views)
I want to generate the median of a variable, and I understand that PROC SQL is only capable of calc median horizontally and not vertically. Now my question is, if I use proc means, is there a way to output the median to a macro variable? I'm hoping to find a more code-efficient way instead of having to write another datastep to call symput to a macro variable.
my current code:
proc means data=tmp_aaa noprint;
var pvol;
output out=tmp_median_aaa median=tmp_median_aaa;
;
run;
data tmp_median_aaa;
set tmp_median_aaa;
call symput('tmp_median_aaa', tmp_median_aaa);
run;
my current code:
proc means data=tmp_aaa noprint;
var pvol;
output out=tmp_median_aaa median=tmp_median_aaa;
;
run;
data tmp_median_aaa;
set tmp_median_aaa;
call symput('tmp_median_aaa', tmp_median_aaa);
run;
3 REPLIES 3
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
Hello.
From my knowledge this is the most efficient way of doing this.
You could use the SELECT INTO: feature of PROC SQL to populate a macro var after doing the necessary aggregation. But as you already pointed out, there is no support for MEDIAN(X) summary functions (vertically), just the common datastep MEDIAN function (horizontally).
Cheers from Portugal.
Daniel Santos @ www.cgd.pt.
From my knowledge this is the most efficient way of doing this.
You could use the SELECT INTO: feature of PROC SQL to populate a macro var after doing the necessary aggregation. But as you already pointed out, there is no support for MEDIAN(X) summary functions (vertically), just the common datastep MEDIAN function (horizontally).
Cheers from Portugal.
Daniel Santos @ www.cgd.pt.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I don't think there is a more "code efficient" way than using call symput. This is the way I do it.
- Mark as New
- Bookmark
- Subscribe
- Mute
- RSS Feed
- Permalink
- Report Inappropriate Content
I agree with the other posters. A very (very) slight improvement would be to avoid outputting the median data set again, as in:
data _null_;
set tmp_median_aaa;
call symputx('tmp_median_aaa', tmp_median_aaa);
run;
data _null_;
set tmp_median_aaa;
call symputx('tmp_median_aaa', tmp_median_aaa);
run;