Date: Wed, 7 Feb 2007 14:51:16 -0600
Reply-To: Gregg Snell <sas-l@DATASAVANTCONSULTING.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Gregg Snell <sas-l@DATASAVANTCONSULTING.COM>
Subject: Re: Teradata Vs. SAS
In-Reply-To: <200702071438.l17Bk0kP008170@mailgw.cc.uga.edu>
Content-type: text/plain; charset=us-ascii
Mike,
You will undoubtedly see a ppt slide with the following:
Extracting Data from Teradata
Example: Frequency Analysis on 3 million row table, on column
Proc Freq w/ SAS/ACCESS for ODBC: 10 minutes, 21 seconds
Proc Freq w/ SAS/ACCESS for Teradata: 6 minutes, 3 seconds
Teradata SQL: 15 seconds
What they don't tell you is that ALL of the time increase for using SAS with
Teradata is in getting the data out of Teradata and into SAS. The Teradata
extract is fast, but the pipe between the Teradata box and SAS box is
probably fairly small.
Here is a log from a freq I did on data already in SAS:
55 proc freq data=pfdec.post_migration_content;
56 tables tdr06;
57 run;
NOTE: Writing HTML Body file: sashtml1.htm
NOTE: There were 6890129 observations read from the data set
PFDEC.POST_MIGRATION_CONTENT.
NOTE: The PROCEDURE FREQ printed page 1.
NOTE: PROCEDURE FREQ used (Total process time):
real time 5.24 seconds
cpu time 5.10 seconds
Ouch! That's gotta hurt! More than twice the rows but still 3x faster than
Teradata.
So, when evaluating Teradata/SQL vs. SAS, be sure the data does not have to
be ported across systems.
Also, when joining a SAS table with a table in Teradata it can be EXTREMELY
slow. So much so, that the quickest method is generally to upload the SAS
table (or at least the keys) to a temp Teradata table, join that to the
table you want, then pull that entire table down to SAS and continue
working.
Good luck,
Gregg Snell
-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Mike
Durbin
Sent: Wednesday, February 07, 2007 8:38 AM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Teradata Vs. SAS
All,
I just recieved an invitation from Teradata to attend a session on how
to use SQL to replace SAS code and run much faster. Also they are touting
advanced analytics within their database that would also replace SAS. If
anyone has experience with both products I'd love to hear the pros and
cons of what Teradata is promoting.
Thanks,
Mike Durbin