dynamic datasets using set

[SRINIVAS_N]

%let start_year=2013;

data time_period(drop=year month);
do year=&start_year to year(today());

do month=1 to 12;
tp=input(compress(year||put(month, z2.)), 6.);

if tp>put(today(), yymmn6.) then
delete;
output;
end;
end;
run;

proc sql;
select count(tp)
into : cnt
from time_period;
quit;

proc sql;
select tp
into :pk1 - :pk%sysfunc(compress(&cnt))
from time_period
order by tp;
quit;

%macro datasets;
%do date=1 %to &cnt;
data dsn_&&pk&date;
x=10;
run;
%end;
%mend Datasets;

%datasets

the datasets will create as

dsn_201301

dsn_201302

dsn_201303

dsn_201304

dsn_201305

.

.

.till today (dsn_201810)

what ever datasets are creating all datasets has to append a one dataset dynamically.

like 

data one;

set dsn_201301 dsn_201302 dsn_201303 ...............dsn_201810

[Reeza]

Can you explain the overall issue? There are ways to shortcut reference data set lists, such as colon or - operator. But it depends on your requirements.

ie SET dsn_201301-dsn_201305; is perfectly valid.

[SRINIVAS_N]

In the set statement datasets should be dynamic

[Kurt_Bremser]

---

@SRINIVAS_N wrote:
In the set statement datasets should be dynamic

---

That's exactly what the macro is for. It creates SAS dates in monthly sequence. Just use the proper format and build the dataset names with the resulting strings.

[ballardw]

I'm not sure if the question is creating a bunch of data sets or combining them.

If the sets already exist then no need to create a set statement listing every single one unless you have some that you need to exclude for some reason.

data one;

   set dsn_201301 -  dsn_201810;

run;

should work.

So if you need need specific start and end "dates" you only need that, start and end.

[SRINIVAS_N]

I should not write statically it should be dynamic

[Reeza]

https://support.sas.com/kb/37/150.html

[Astounding]

Is this close enough to being dynamic?

set dsn_20: ;

It takes all data set names that begin with "dsn_20"