Problem with Put and missing data

[BruceBrad]

Why doesn't the following code work? When I run it, I get 20 commas in the log (followed by the value of i, 101 on the next line), but I am expecting five rows of 20 commas. If I change it to v(i)=1, it works as expected, with multiple rows of 1 followed by a comma.  [This is abstracted from a large program where missing values seem to messing up my output]

data _null_;
array v(*) v1-v100;
do i=1 to dim(v);
  v(i)=.;
end;
file log dsd ls=20 dlm="," ;
put (_all_) (:);
run;

[gamotte · Accepted solution]

It seems like a bug to me.

As a workaround, you could use a codification for missing values, for instance :

option missing="M";

[BruceBrad]

I'm trying to read the data back into SAS using

input (_all_)( : );

(After first defining the variable list and informats using attrib statements).

So, to use your work around I have to add

missing M;

prior to the input statement. This works, though the missings are now read in as .M rather than generic missing (.). I then need to follow up with array loop to recode them to my preferred generic missing.

Thanks

PS I've sent this on to tech support.

[gamotte]

Hello,

You can use a user defined informat to interpret M's as missing values.

proc format;
    invalue v
	    "M"=.
	    other=best.
    ;
run;

option missing=".";
data test;    
    informat v1-v100 v. i best.;
    infile log1 dlm="," dsd;
	input (_all_) (:);
run;