topic Re: first.gvkey in SAS Programming

first.gvkey

abdulla — Wed, 30 Sep 2020 19:02:08 GMT

gvkey	fyear	at	want
1001	2001	12.3	.
1001	2002	13.6	12.3
1001	2003	12.6	13.6
1001	2004	11.9	12.6
1002	2001	13.4	.
1002	2002	17.7	13.4
1002	2003	18.4	17.7
1002	2004	15.1	18.4

data want;

set have;

by gvkey fyear;

if first.gvkey then want=.;

else want=lag(at);

run;

My data is already sorted. I am not getting the expected result. What is the mistake here?

Re: first.gvkey

mklangley — Wed, 30 Sep 2020 19:05:15 GMT

@abdulla Can you include both your input data and your expected result?

Re: first.gvkey

Reeza — Wed, 30 Sep 2020 19:07:54 GMT

Lag cannot be calculated conditionally, so calculate it all the time and then set it to missing AFTER.

want = lag(at);
if first.gvkey then call missing(want);

Re: first.gvkey

Kurt_Bremser — Wed, 30 Sep 2020 19:08:05 GMT

Never use LAG conditionally (unless you really know what you are doing). Slightly change your code:

data want;
set have;
by gvkey fyear;
want = lag(at);
if first.gvkey then want = .;
run;

This is necessary because LAG will only put a value into the queue when it is called.

Re: first.gvkey

abdulla — Wed, 30 Sep 2020 19:08:25 GMT

The first three columns are my data and the fourth column is my expected result.

Re: first.gvkey

abdulla — Wed, 30 Sep 2020 19:13:23 GMT

Thank you