Date: Mon, 9 Jun 2008 12:06:23 -0400
Reply-To: Hari Nath <hari_s_nath@YAHOO.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Hari Nath <hari_s_nath@YAHOO.COM>
Subject: Re: Oracle to SAS to DB2
Content-Type: text/plain; charset=ISO-8859-1
Thanks Howard and Gerhard for your suggestions....Will give a try...
On Wed, 4 Jun 2008 05:43:55 -0400, Gerhard Hellriegel
>and note: physically that is the same. Both are datasets on a harddisk,
>depending on the OS you use. Important for sure is the kind of that
>harddisks: local ones are normally faster than remote ones, RAID systems
>can (!) be faster than no-RAID, depending on the RAID-level, new harddisks
>are some´times faster than old ones (not because they get slower, but
>because new technology might be faster. Not always, e.g. the new
>electronic disks are faster in reading, but slower in writing).
>To find out the best way: have a look at all storage which is available
>and big enough, then try it out.
>By the way: it could be a good idea to take a local disk, but NOT that one
>where SAS and all the other programs reside. Install a dedicated harddisk
>for your data-files!
>All is different (and good!), if you are on a zOS-mainframe: try to use
>hiperspace in that case! That might be really fast, depending on where the
>hiperspace is realized and how many others are using it at the same time.
>On Tue, 3 Jun 2008 19:26:31 -0400, Howard Schreier <hs AT dc-sug DOT org>
>>On Sat, 31 May 2008 13:14:39 -0400, Hari Nath <hari_s_nath@YAHOO.COM>
>>>Is there a difference in performance between loading a db2 table from a
>>>temporary sas dataset versus a permanent sas dataset.
>>>I am running a process in Oracle wherein the output is stored in a sas
>>>dataset and then this data is fed back into DB2. My question is whether
>>>there will be an improvement in performance if the created sas dataset
>>>stored in a permanent library other than work. Am using sas 8.2
>>>Thanks for your reply in advance!!
>>I think it depends less on temporary vs. permanent and more on where the
>>libraries live on the network. Try to create the intermediate data set on
>>the box where SAS is running.