How can I add some missing observations in a dataset?

[Shayan2012]

Hi all,

I have this data set on firm-years with their location addresses from 1990-2005. The original data set has not the address for all of the years, so there is no observation included for that specific year. However, I want to add those firm-years to the data set in an efficinet way.

So, this is my current data set:

Firm	Year	Address
x	1990	xxx
x	1991	xxx
x	1994	xxx
x	1996	yyy

and I want to convert in into:

Header 1	Header 2	Header 3
x	1990	xxx
x	1991	xxx
x	1992	---
x	1993	---
x	1994	xxx
x	1995	---
x	1996	yyy

Can anyone please help me how to do that efficiently?

Thanks a lot!

[Haikuo · Accepted solution]

This I believe has been asked and answered many times, so do a search you will learn more solutions, the following is just one of them.

data have;

input Firm$ Year Address$;

cards;

x 1990 xxx

x 1991 xxx

x 1994 xxx

x 1996 yyy

;

data want;

  merge have have(keep=year firm rename=(year=_y firm=_f) firstobs=2);

output;

if firm=_f then do year=year+1 to _y-1;

call missing (address);

  output;

  end;

drop _:;

run;

Haikuo

Last version was not rigorously tested, hence the update.