## Summary file for discrete value

Solved
Super Contributor
Posts: 506

# Summary file for discrete value

[ Edited ]

Hi Everyone,

I already have a file in which each variable has noduplicate value.

I want to create a summary file that show the number of value for each variable in a different format as below

data have;
input a b c;
cards;
1 1 2
-2 2 .
5 3 .
6 . .
run;

data want has 2 column: variable_name and number

a 4  meaning there are 4 values with variable a

b 3

c 1

Since there are hundreds of variables, I dont know how to make it work efficiently.

Any help is very much appreciated.

HHC

****************************************************

proc summary data=have noprint;
var _numeric_;
output out = a n=; run;

proc transpose data=a out=a1;
var _numeric_;
run;

Proc means data=have stackods N NMISS;
var _numeric_;
ods output summary=want;
run;

Accepted Solutions
Solution
‎01-20-2017 10:08 AM
Super User
Posts: 6,933

## Re: Summary file for discrete value

Since there are no duplicate values, this is really just a count of how many non-missing values exist.  This will get it horizontally:

proc summary data=have;

var _numeric_;

output out=stats (drop=_type_ _freq_) n=;

run;

At this point you could either use the data set STATS, or you could transpose it to get what you described as the WANT data set:

proc transpose data=stats out=want (rename=(col1=number));

var _numeric_;

run;

All Replies
Solution
‎01-20-2017 10:08 AM
Super User
Posts: 6,933

## Re: Summary file for discrete value

Since there are no duplicate values, this is really just a count of how many non-missing values exist.  This will get it horizontally:

proc summary data=have;

var _numeric_;

output out=stats (drop=_type_ _freq_) n=;

run;

At this point you could either use the data set STATS, or you could transpose it to get what you described as the WANT data set:

proc transpose data=stats out=want (rename=(col1=number));

var _numeric_;

run;

Super User
Posts: 24,003

## Re: Summary file for discrete value

Proc means with N statistic should calculate this.

Ods select none;

Proc means data=SASHELP.class stackods N NMISS;

var _numeric_;

ods output summary=want;

run;

ods select all;

☑ This topic is solved.