topic Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum in SAS Programming

Use "firstobs" in DATA step to create lag for a single column instead of all the columns.

Saurabh_Rana — Mon, 30 Mar 2020 10:30:23 GMT

I use firstobs to create lag and merge it with existing data, but every time I have to rename all the columns as C1=_C1, C2=_C2 so that I can drop them easily since they are of no use to me.This is easy if the table has only 4-5 columns but I am working on datasets which have more than 50 columns.So I have to rename them all manually since I don't know how to use macros.

Is there a way through which I can create lag for a specific column of a table?

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

Ksharp — Mon, 30 Mar 2020 10:37:22 GMT

IF your variables like C1 C2 .C10 .
could try :

data want;
merge have have( rename=(C1-C10 = _C1-_C10) );
run;

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

Saurabh_Rana — Mon, 30 Mar 2020 10:38:44 GMT

No that was just a example I used to explain the question.

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

Saurabh_Rana — Mon, 30 Mar 2020 10:42:24 GMT

Also is there a way to create lead instead of lag using firstobs or some other function in SAS.

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

Ksharp — Mon, 30 Mar 2020 10:45:50 GMT

Here is . If you don't mind.

data have;
 set sashelp.class;
run;


proc transpose data=have(obs=0) out=temp;
var _all_;
run;
proc sql noprint;
select cats(_name_,'=_',_name_) into :rename separated by ' ' from temp;
quit;
data want;
 merge have have(firstobs=2 rename=(&rename));
run;

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

andreas_lds — Mon, 30 Mar 2020 10:47:06 GMT

@Saurabh_Rana wrote:
No that was just a example I used to explain the question.

Then please post an excerpt of the data you have as usable data step, so that we don't have to guess what you have. And please show what you expect as result.

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

Saurabh_Rana — Mon, 30 Mar 2020 10:49:21 GMT

I can't mention the columns but what I am saying is that all the column names are completely different.

Re: Use "firstobs" in DATA step to create lag for a single column instead of all the colum

Kurt_Bremser — Mon, 30 Mar 2020 11:17:56 GMT

You are NOT(!!!) creating a lag, you create a look-ahead.

In the look-ahead table, use a keep= dataset option to keep only the column(s) needed:

data want;
merge
  have
  have (keep=c rename=(c=_c))
;
/* further processing */
run;