[ESS] ESS[SAS]: Submit region starting new SAS session?
P@u|@Thomp@on @end|ng |rom S@n|ordHe@|th@org
Thu Feb 12 23:34:19 CET 2015
I have a few suggestions and will send them to you tomorrow. Amazingly enough, my talk at SGF this year is about designing workflow management tools in SAS. I have a document which discusses this subject as well.
From: ESS-help [mailto:ess-help-bounces using r-project.org] On Behalf Of Erin Jonaitis
Sent: Thursday, February 12, 2015 4:11 PM
Subject: Re: [ESS] ESS[SAS]: Submit region starting new SAS session?
Hi Rodney and Paul,
Thanks for the suggestions. It looks like I'm in version 5.11 -- clearly a few versions behind. Upgrading is not in the cards, I'm afraid, though I appreciate that that makes it hard for others to help.
> Well, they are two totally different beasts. In R, you can easily run
> an analysis with a one-liner. In SAS, you need at least 3 lines to do
> anything. Also, SAS does not have the concept of objects so there are
> no temporary objects to manipulate other than data sets (and
> occasionally, catalogs) which are not really objects at all.
Most of my workflow in SAS has been about data management -- understanding how the data tables relate to one another, making sure I've merged them correctly, finding unexpected things (outliers, missingness, illegal values), and occasionally using procs to create secondary datasets of parameter estimates etc. So the loss of the ability to keep datasets around for interactive poking and prodding is actually a pretty big stumbling block for me -- they're not as powerful as the data structures you get in R, but they're most of why I use SAS at all.
I wonder how it works in Windows? The library names, temporary datasets, etc. do persist on Windows until the program sas.exe is shut down. The version of SAS I have access to here is the same one I started programming with in Windows, so I know that's not principally a function of code age.
I liked Paul's suggestion to define things in macros. It doesn't solve the problem of having to rerun all the code every time, but it does hide the data setup so it's not cluttering the remainder of my program. It likely has some other virtues too. Combining that with judicious use of temporary directories in which to save permanent sas datasets while I'm looking at them -- and/or possibly moving some of my data exploration to R -- I suspect I'll be able to reconstruct most of the workflow I'm used to. It'll just take some doing.
Thanks again for walking me through this. Lots to learn!
ESS-help using r-project.org mailing list
Confidentiality Notice: This e-mail message, including any attachments,
is for the sole use of the intended recipient(s) and may contain
privileged and confidential information. Any unauthorized review, use,
disclosure or distribution is prohibited. If you are not the intended
recipient, please contact the sender by reply e-mail and destroy
all copies of the original message.
More information about the ESS-help