Find the non-missing value

[Babloo]

I've a dataset like below.

A    B

1    .

.     2

3    .

.     4

5    .

I need the dataset like below. Basically I need the non-missing values for each observation. May I request someone to guide me to accomplish this task via Proc SQL?

A    B   C

1    .     1

.     2    2

3    .     3

.     4    4

5    .     5

[sbxkoenk]

Like this.

The MAX function ignores null values and SAS missing values. So the missing is completely ignored. Thus: max does not return 5 because 5 > . , the missing is just not considered.

data ab;
input a b;
datalines;
1    .
.    2
3    .
.    4
5    .
;
run;

PROC SQL noprint;
 create table abc as
 select * , max(a,b) as c  
 from ab;
QUIT;
/* end of program */

Koen

[Reeza]

I would recommend the coalesce() function which takes the first non missing value in a series of variables specified.

Coalesce(a, b) as C