|
On Thu, 9 Nov 2006 11:02:09 -0500, Steven Lolay <stevenlolay@IF.COM> wrote:
>Hi
>
>Is it possible for you to produce a datset which has 2 records for each
>category ie a Y and No with the code below:
>
>DATA dummy_1;
>INPUT match 1. cat $20.;
>
>DATALINES;
>1Fixed
>1Tracker
>1Offset not Active
>1Offset Active
>run;
>proc sort data=dummy_1;by match;;run;
>
>
>DATA dummy_2;
>INPUT match 1. erc $1.;
>DATALINES;
>1N
>1Y
>run;
>proc sort data=dummy_2;by match;run;
>
>
>data dummy;
>
>merge dummy_1 dummy_2;
>
>by match;
>
>run;
>
>
>I think it's called a cartesian join but could be wrong!!!
>
>Could you please tick the box to send me a reply as in a rush.
>
>Thanks
It's done most easily with SQL, and unless you really have additional groups
you do not need the MATCH columns.
DATA dummy_1;
INPUT cat $20.;
DATALINES;
Fixed
Tracker
Offset not Active
Offset Active
;
DATA dummy_2;
INPUT erc : $1.;
DATALINES;
N
Y
;
proc sql;
create table dummy as
select * from dummy_1 cross join dummy_2
order by cat, erc desc;
quit;
Result:
cat erc
Fixed Y
Fixed N
Offset Active Y
Offset Active N
Offset not Active Y
Offset not Active N
Tracker Y
Tracker N
| 
