number of missing responses for each observation

[code_blooded]

Hi! 

I am looking to determine the number of missing values for each observation (rows) in my dataset which has both numeric and character variables. it is a fairly large dataset >500 variables and has a mix of both character and numeric variables. I tried using cmiss but it said that the variables had to be either all character or all numeric.

Any insights??

[PaigeMiller · Accepted solution]

number_missing = cmiss(of _character_) + nmiss(of _numeric_);

[code_blooded]

Thanks Paige. This worked however I am getting one extra count for the missing values. ANy idea what might have cause this? 

Below is the code that I have been using:

data miss;
set test;
nmiss = cmiss(of _character_) + nmiss(of _numeric_);
run;

Thanks for the help!

[Patrick]

@code_blooded 

It works for me. Before investigating your code may be execute below and verify that this also works in your environment and for your SAS version.

data test;
  length ca cb $1 na nb 8;
  call missing(of _all_);
  na=1;
  n_missing=cmiss(ca,cb,na,nb);
  output;
  stop;
run;
proc print;
run;