proc means output

[Dorota_Jarosz]

Expanding on Linlin's ideas and assuming SAS version 9.1 and no SAS/Access to PC Files:

proc means data = sashelp.class noprint;
var age ;
    OUTPUT OUT=dat1 (drop=_:) q1=q1 q3=q3 qrange=qrange ;
run;
proc means data = sashelp.class noprint;
var height ;
    OUTPUT OUT=dat2 (drop=_:) q1=q1 q3=q3 qrange=qrange ;
run;
proc means data = sashelp.class noprint;
var weight ;
    OUTPUT OUT=dat3 (drop=_:) q1=q1 q3=q3 qrange=qrange ;
run;
data want;
length item $ 10;
  set dat1;
  item="age";
  output;
  set dat2;
  item="height";
  output;
  set dat3;
  item="weight";
  output;
run;
proc print;run;
*output:
Obs    item       q1      q3     qrange

1     age       12.0    15.0       3
2     height    57.5    66.5       9
3     weight    84.0    112.5     28.5
;

*old fashion CSV output to flat file;
filename csv "C:\temp\Quartiles.csv";
data _null_;
set want;
file csv;
if _n_=1 then
put @1 "Variable,"
    @12 "q1,"
    @22 "q3,"
    @32 "qrange"
    ;
put @1  item      ","
    @12 q1 best8. ","
    @22 q3 best8. ","
    @32 qrange best8.
    ;
run;

[MikeZdeb]

hi ... rather than use a data step to assemble the data after MEANS, how about using a data step prior to SUMMARY and running it once ...

data class (keep=variable value) / view=class;

set sashelp.class;

array nm(*) _numeric_;

do j=1 to dim(nm);

   variable = vname(nm(j));

   value = nm(j);

   output;

end;

run;

proc summary data=class nway;

class variable;

var value;

output out=stats (drop=_:) q1=q1 q3=q3 qrange=qrange;

run;

variable     q1       q3     qrange

Age        12.0     15.0      3.0

Height     57.5     66.5      9.0

Weight     84.0    112.5     28.5

* create CSV file without ACCESS to PC Files module;

ods csv file='z:\stats.csv';

proc print data=stats noobs;

run;

ods csv close;

[Dorota_Jarosz]

Thank you, MIke. I really like your solution. My mind is still in the mainframe and flat file mode.

[Hawkeye]

I ended up geting sas 9.3 so I was able to use the stackods option, thank you all for the help.