LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (February 2004, week 1)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:   Mon, 2 Feb 2004 18:38:49 -0500
Reply-To:   Sigurd Hermansen <HERMANS1@WESTAT.COM>
Sender:   "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:   Sigurd Hermansen <HERMANS1@WESTAT.COM>
Subject:   Re: SQL speed
Comments:   To: Alexander Cavallo <acavallo@LEXECON.COM>
Content-Type:   text/plain

Alexander: SAS SQL relies on the sort routines being used by the SAS System. Some installations of the SAS System use SyncSort. So far as I can tell, SyncSort generally works faster than other sort routines, although in many applications the differences are small. Sig

-----Original Message----- From: Alexander Cavallo [mailto:acavallo@LEXECON.COM] Sent: Monday, February 02, 2004 6:07 PM To: SAS-L@LISTSERV.UGA.EDU Subject: SQL speed


I am new to using SQL. From reading SAS-L, I thought that SQL sorts faster than SYNCSORT. In the example below, SYNCSORT is faster. Could you explain to me how big a dataset needs to be before SQL is faster?


--Alex Cavallo Lexecon, Inc.

48 proc sql noprint; 49 create table temp.sql as ( 50 select distinct emp_id, date 51 from hrdata.all ) 52 order by emp_id, date 53 ; NOTE: Table TEMP.SQL created, with 179291 rows and 2 columns.

54 quit; NOTE: The PROCEDURE SQL used 2.22 CPU seconds and 10730K.

55 56 proc sort nodupkey data=hrdata.all out=temp.test(keep=emp_id date); 57 by emp_id date; 58 run;

NOTE: WER750I End PROC SYNCSORT. R2.2D+ NOTE: There were 179291 observations read from the data set HRDATA.ALL. NOTE: The data set TEMP.TEST has 179291 observations and 2 variables. NOTE: The PROCEDURE SORT used 1.88 CPU seconds and 10907K.

Back to: Top of message | Previous page | Main SAS-L page