Date:         Fri, 1 Aug 2008 15:31:00 -0400
Reply-To:     Chang Chung <chang_y_chung@HOTMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Chang Chung <chang_y_chung@HOTMAIL.COM>
Subject:      Re: SQL to SAS
Comments: To: Mary H <mlhoward@AVALON.NET>

On Fri, 1 Aug 2008 13:12:54 -0500, Mary <mlhoward@AVALON.NET> wrote:

>A preferred approach (unless you really need it in SAS, is to assign the
libname like Charles indicated, then use it directly just as you would a SAS
data set:
>
>proc freq data=mydblib.table_name;
...
hi, mary,

i read several of your sas-l postings recently and concluded that you must
be a computer hardware sales person.

First, you said that it will be OK to run a job that will bloat a dataset
with several thousands observations to one with 500+ million observations
(OK. that was my code, but i did warn the OP; and Ian came up with a better
approach.)

Then you suggested a macro that calls proc sql 3,000 times to solve a
problem that OP said that s/he did it with a single invocation of the proc
sql (and proc content).

Finally, you are saying that it is a "preferred" approach to pull out data
from a db system over and over, every time you need to call a proc.

of course, these are "good" practices to slow down a computer, to exhaust
available memories, and to clog network pipelines. i guess that your sales
are down and you have to somehow convince people that they need to buy
newer, faster, bigger hardware pieces, i guess. :-)

cheers,
chang