Date:         Fri, 21 Jun 2002 06:24:52 -0700
Reply-To:     shiling zhang <shiling99@YAHOO.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         shiling zhang <shiling99@YAHOO.COM>
Organization: http://groups.google.com/
Subject:      Re: looping match merge?
Content-Type: text/plain; charset=ISO-8859-1

SQL can create this type of product(Cartesian Product) easily. Please
take a look the follow example. It blows up really quick. The total
obs will be NOBS(dataset t1) X NOBS (dataset t2).  You also can create
Cartesian Product within a key group too.

HTH

data t1 t2;
  do x=1,2,3;
    output t1;
  end;
  do y=4,5,6;
    output t2;
  end;
run;

proc sql;
  select t1.x , t2.y
  from t1 , t2
;
quit;

Arthur Ellen <arte@panix.com> wrote in message news:<aetp6g$phd$1@reader1.panix.com>...
> I'm trying something like this example,
> where I want to perform a match merge,
> but go back and sequentially go through
> the smaller file.
>
> DATA A;
>   INPUT X;
>   MERGEVAR=99;
>   CARDS;
>   1
>   2
>   3
>   4
>   5
>   ;
> DATA B;
>   INPUT Y;
>   MERGEVAR=99;
>   CARDS;
>   5
>   10
>   15
>   ;
> DATA C;
>   MERGE A B;
>      BY MERGEVAR;
>   /* I want to be able to just use
>      _N_=1 for Data B then go
>      select the 2nd until Data B completes
>   */
>   ARRAY Z(3);
>   Z(1)=X+Y;
>   Z(2)=X+Y;
>   Z(3)=X+Y;
> PROC PRINT;
> RUN;
>
> What I'd like to get
>
>                        X     MERGEVAR   Y     Z1    Z2    Z3
>
>                  1     1       99       15     6    11    16
>                  2     2       99       15     7    12    17
>                  3     3       99       15     8    13    18
>                  4     4       99       15     9    14    19
>                  5     5       99       15    10    15    20
>
> art ellen
> arte@panix.com