topic sas entries manipulation in SAS Procedures

sas entries manipulation

jkf91 — Sun, 02 Oct 2011 18:52:33 GMT

save i have the following data:

Observation	x	y
1	1	2
2	3
3	2

as you can see, column y is not complete. BUT, all rows in column y should take the same value:

Observation	x	y
1	1	2
2	3	2
3	2	2

Two Questions:

First, I want to know how I can do that on SAS.

Next, I need want to keep observations such that x>=y. How can I do that?

Thanks,

sas entries manipulation

art297 — Sun, 02 Oct 2011 18:59:31 GMT

One way would be:

data want (drop=_:);

set have (rename=(y=_y));

retain y;

if _n_ eq 1 then y=_y;

if x gt y;

run;

sas entries manipulation

jkf91 — Sun, 02 Oct 2011 21:19:19 GMT

thanks a lot! that solved the problem!

just curious though....

what are these two doing?

drop=_:

rename=(y=_y)

thanks,

sas entries manipulation

art297 — Sun, 02 Oct 2011 21:29:53 GMT

Just a couple of shortcuts. The rename statement changes the name of variable y, when the file is set, so that it will be seen to be variable _y; I did it that way so that the later missing values of y wouldn't override the desired value from the first record.

The drop statement, then, just uses the semicolon as a wild card. It says, in essence, to drop any variable that starts with an underscore (i.e., variable _y).

The semi colon has lots of under used useage in SAS as stated, nicely, in:

http://www2.sas.com/ proceedings/sugi29/054-29.pdf