topic Re: identify not missing observations in SAS Programming

identify not missing observations

janie99 — Thu, 05 Jun 2014 08:32:23 GMT

I want to create a new variable to show which rows have no missing observations. Rows with no missing variables are coded as 1 in the notmissing variable and 0 otherwise

i.e.

Obs	Var1	Var2	Var3	notmissing
1	2	3	.	0
2	3	2	4	1
3	.	3	1	0
4	7	4	2	1

The code that I have is not working. This is what I have tried:

data work.dataset 
set work.dataset
if var1^=. and var2^=. var3^=. then notmissing=1;
else notmissing=0;
run;

Re: identify not missing observations

RW9 — Thu, 05 Jun 2014 08:58:14 GMT

Hi,

Change you code to (as you are missing an and:

if var1 ne . and var2 ne . and var3 ne . then notmissing=1;

else notmissing=0;

Also, if you are using SAS 9.2 then you can use the CMISS() function and if > 0 then 1 else 0. There are other ways too, check out arrays if you have lots of variables to check, then you could use a do loop to go over it.

data have;

attrib var1 var2 var3 notmissing format=best.;

var1=2; var2=3; var3=.; output;

var1=3; var2=2; var3=4; output;

run;

data want (drop=i);

set have;

array var{3};

notmissing=1; /* Assume not missing */

do i=1 to 3;

if var{i} = . then notmissing=0; /* if any are missing then set to missing */

end;

run;

The benefit of the above becomes clear when you have 10 or 20 var variables, as you just need to increase the loop size.

Re: identify not missing observations

jakarman — Thu, 05 Jun 2014 10:15:49 GMT

SAS(R) 9.4 Functions and CALL Routines: Reference, Second Edition (NMISS function)

Using with variable lists you will get easy code. Just all variables being non-missing will give a value of 0 and all with some missings above.

Re: identify not missing observations

Dalveer — Thu, 05 Jun 2014 10:22:07 GMT

hi,

the code which you have written is perfectly alright just you missed out one 'AND' operator between var2,var3.

corrected statement:-

if var1 ne . and var2 ne . and var3 ne . then notmissing=1;

Re: identify not missing observations

slchen — Thu, 05 Jun 2014 10:24:45 GMT

data have;

input id var1-var3;

cards;

1 2 3 .

2 3 2 4

3 . 3 1

4 7 4 2

;

run;

data want;

set have;

missing_flag=(nmiss(of var1-var3)=0);

run;

Re: identify not missing observations

stat_sas — Thu, 05 Jun 2014 10:31:25 GMT

data want;

set have;

flag=nmiss(of var1-var3);

run;

Re: identify not missing observations

janie99 — Wed, 11 Jun 2014 09:19:37 GMT

Thanks all for your help. For some reason the "and" function wasn't working so just did each variable separately and added them together

Re: identify not missing observations

Chrishi — Thu, 12 Jun 2014 06:23:16 GMT

it should be 'or' not 'and' for the result you need.

Sudeer.

Re: identify not missing observations

Vish33 — Thu, 12 Jun 2014 09:38:09 GMT

Hi,,

Hope OR should work instead of AND ;

data work.dataset;

set work.dataset;

if Var1 = . OR Var2 = . OR Var3=. then notmissing = 0 ;

else notmissing = 1;

run;