Can I include the dataset name in a query (proc sql or data step) without hardcoding it

[DritanB]

Suppose I have two datasets named ds1, ds2 that have each 4 columns (their structure is the same):

ds1:

col1 | col2 | col3 | col4

================

122 | aaa | a1a1 | b1b1

124 | a1a | a2    | b2bb

235 | aaa |  e34 | r45

ds2:

col1 | col2 | col3 | col4

================

224 | baa | a2a1 | b1b2

124 | a1a | a2    | b2bb

547 | aaa |  e54 | v45

I am trying to write some code that will create a  dataset containing col1 and col2 and a new column that would contain the dataset name.

The condition would be col2='aaa'

And the new dataset would look like this:

col1 | col2 | name

=============

122 | aaa | ds1

235 | aaa | ds1

547 | aaa | ds2

I wouldn't want to hardcode the name of the datasets in the new query.

Any suggestion would be greatly apprecited.

[ChrisNZ · Accepted solution]

Something like this?

data WANT;
  set DS1 DS2 indsname=INDS;
  where COL2='aaa';
  NAME=INDS;
  keep COL1 COL2 NAME;
run;

[DritanB]

Thank you!