SAS--calculating mean by multiple groups

问题

Supposing a dataset as following:

id    date         var1
001   20170101    1
001   20170101    2 
001   20170101    3
001   20170102    1
001   20170102    2
002   20170101    1
002   20170101    2
002   20170102    1
002   20170102    2

I to calculate the mean for each id in each date through following code.

proc summary data=HAVE nway;
class id date;
var var1 ;
output out=WANT(drop=_:) mean=mean std=std;
run;

However, the WANT only represents date, mean, and std, but does not contain ID. How could I solve this problem?

回答1:

I wasn't able to reproduce your issue. After passing

   data have;
     informat date yymmdd8.;
     input id date  var1;

   datalines;
   001 20170101 1
   001   20170101    2 
   001   20170101    3
   001   20170102    1
   001   20170102    2
   002   20170101    1
   002   20170101    2
   002   20170102    1
   002   20170102    2
   ;
   run;

   proc summary data=HAVE nway;
     class id date;
     var var1 ;
     output out=WANT(drop=_:) mean=mean std=std;
   run;

I got the HAVE dataset like like you showed but the resulting dataset WANT includes both class variables. Maybe your actual id variable is prefixed with underscore and dies with the drop=_: ?

来源：https://stackoverflow.com/questions/47444022/sas-calculating-mean-by-multiple-groups