Date:         Tue, 6 Sep 2005 15:31:33 -0400
Reply-To:     Sigurd Hermansen <HERMANS1@WESTAT.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Sigurd Hermansen <HERMANS1@WESTAT.COM>
Subject:      Re: limiting sql rows by calculated monotonic()
Comments: To: Nina Harris <sas@mailinator.com>
Content-Type: text/plain; charset="US-ASCII"

Nina:
Selecting at random about a thousand rows from the first 100,000 would
give you a better look at the data and support more a reliable analysis.
I'd likely go with a Data step view (IF _N_<=100000;) that keeps rows
'IF ranuni(11111) < 0.1;'. You can continue to use a Data step view of a
SAS/Access engine database source.
Sig

-----Original Message-----
From: owner-sas-l@listserv.uga.edu [mailto:owner-sas-l@listserv.uga.edu]
On Behalf Of Nina Harris
Sent: Tuesday, September 06, 2005 2:05 PM
To: sas-l@uga.edu
Subject: limiting sql rows by calculated monotonic()


are there better ways of doing this?

the goal is to "look and analyze" at the data coming from a hundred
million row table first before running a full throttle query

* this is here just to generate some sample data
 in real life this would not exist as the sql connection would provide
the data; data a;do x=1000 to 500 by -1;output;end;


proc sql;
 create table b as select *, monotonic() as _m_ from a
 where calculated _m_ < 10
 order by x
;
quit;
data _null_;set b;put(_all_)(=);run;