CODEWITHSUNDEEP
sas
Atreyi Datta
Atreyi Datta

Reputation: 1

SAS Missing Values Findings

I am working on a SAS Dataset which has missing values.
I can identify whether a particular variable has missing values using IS NULL/IS MISSING operator.
Is there any alternative way, through which I can identify which variables have missing values in one shot.

Thanks in Advance

Upvotes: 0

Views: 1826

Answers (4)

SS_SS
SS_SS

Reputation: 43

Two options (in addition to Peter Slezák's) I can suggest are : - Use proc means with nmiss

proc means data = ___ n nmiss;
var _numeric_;
run;
  • In SAS Enterprise Guide, there is a characterize data task - this helps profile character variables too. (Under the hood, it is a combination of various procs, but is an easy to use option).

Hope this helps, regards, Sundaresh

Upvotes: 2

Peter Slezák
Peter Slezák

Reputation: 11

You can use proc freq table statement with missing option. It includes missing category if missing values exist. Useful for categorical data.

data example;
      input A Freq;
      datalines;
   1 2
   2 2
   . 2
   ;

*list variables in tables statement;
proc freq data=example;
    tables A / missing;
run;

You can also use Proc Univariate it creates MissingValues table in ODS by default if any missing values exist. Useful for numeric data.

Upvotes: 1

Tom
Tom

Reputation: 51611

The syntax IS NULL or IS MISSING is limited to use in SQL code (also in WHERE statements or WHERE= dataset options since those essentially use the same parser.)

To test if a value is missing you can also use the MISSING() function. Or compare it to a missing value. So for character variables test if it is equal to all blanks: c=' '. For numeric you can test x=., but you also need to look out for special missing values. So you might test if x <= .z.

To get a quick summary of number of distinct missing values for each variable you could use the NLEVEL option on PROC FREQ. Note it might not work for a large dataset with too many distinct values as the procedure will run out of memory.

Upvotes: 2

Kiran
Kiran

Reputation: 3315

use array and vname to find variable with missing values. If you want rows with missing values use cmiss function.

data have;
  infile datalines missover;
  input id num char $ var $;
datalines;
1 . A C
2 3   D
5 6 B D
;



 /* gives variables with missing values*/

data want1(keep=miss);
set have;
array chars(*) _character_;
array nums(*) _numeric_;

do i=1 to dim(chars);

    if chars(i)=' ' then
        miss=vname(chars(i));

    if nums(i)=. then
        miss=vname(nums(i));
end;

if miss=' ' then
    delete;
run;

  /* gives rows with missing value*/

data want(drop=rows);
set have;
rows=cmiss(of id -- var);

if rows=1;
 run;

Upvotes: 1

Related Questions