Date: Sun, 9 Mar 2008 18:37:13 -0700
Reply-To: Phil <phillip.anderra@GMAIL.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Phil <phillip.anderra@GMAIL.COM>
Subject: Advice on joining large tables wanted.
Content-Type: text/plain; charset=ISO-8859-1
I have two tables, Big & Small that I need to join on the variable P7
and must keep all the records from Big.
Big has 135m rows, 14 variables, all about $10, with the key variable
P7 having 60k distinct, and evenly distributed values.
Small has 170k rows and 2 variables. The key variable P7 is *almost*
Left Join on P7 looked the obvious choice but on my (average) PC it
takes approx 2.5 hours to run.
I've been trawling though this ng and have found suggestions on
improving the run time is situations like this based on a number of
methods, including; indexing, increasing Bufsize & BufNo, using hash
joins etc etc and while I realise that there's no 'One size fits all'
solution would anyone care to give their opinion on a 'best option'
here to help out a slow-moving 9.1 user?