Date:         Wed, 9 Nov 2005 01:12:57 -0800
Reply-To:     Anonymous user <Number_42@CARAMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Anonymous user <Number_42@CARAMAIL.COM>
Organization: http://groups.google.com
Subject:      Re: How to make repeats in merge-by-group to work?
Comments: To: sas-l@uga.edu
In-Reply-To:  <1131467661.492425.249910@g43g2000cwa.googlegroups.com>
Content-Type: text/plain; charset="iso-8859-1"

Thx for your answers:

* "Billings, Tony": your code is really slow for large dataset (in
comparaison with sql)
*  toby dunn: okay, i didn't know that my code "flied in the face of
how the data step works", i won't do it again
*  Dianne Rhodes: thx but i can't find a numeric version on the web
* shilin...@yahoo.com: i already have that, but thx anyway


shiling99@yahoo.com a écrit :

> This merge forms a cartesian product within the id groups in sql. It
> can be simply implemented in SQL as,
>
>
> proc sql;
>   select a.id, fruit, color
>   from one a, two b
>   where a.id=b.id
>   order by 1,2
>   ;
>   quit;
>
> However, the space product grows very fast!.
>
> HTH