Date:         Fri, 9 Feb 2001 16:44:17 -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: SAS Compress vs GZIP
Comments: To: "Wilson-Beavers, Linda" <Linda@CAMBIO.ACOMP.USF.EDU>
Content-Type: text/plain; charset="iso-8859-1"

SAS compress normally refers to options in SAS statements and procedures
that control whether SAS  produces compressed SAS datasets or not.  These
options pertain to SAS datasets only. SAS will decompress a compressed SAS
dataset transparently in SET, MERGE, and other operations that read SAS
datasets.

The GNU compression program, gzip, compresses files.  You can use gzip to
compress a SAS dataset (for example, prior to ftp'ing it from one server to
another) but you will need to decompress it before referring to the SAS
dataset in a SAS program.  (At least I have not found a way yet to have SAS
read a gzip'ped SAS dataset.)

You may find practical uses for both.  The Unix OS makes it easy to pipe
gip'ped data files (not SAS datasets) through a FILENAME statement to an
INPUT statement.  In this case a WHERE condition filters data after
decompression.  In current versions SAS does a better job of handling SAS
dataset compression under MS Windows and Unix.  Incidentally, advocates of
"big iron" computing tell me that transparent compression at the OS level
makes gzip and SAS compress options redundant.  Sig

-----Original Message-----
From: Wilson-Beavers, Linda [mailto:Linda@CAMBIO.ACOMP.USF.EDU]
Sent: Friday, February 09, 2001 3:06 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: SAS Compress vs GZIP


 Does anyone know the differences between using SAS compress and using gzip
to save storage
space?  When would it be more practical to use one or the other?

TIA,
Linda