Thanks Matt and others. I expect I introduced an error in naming variables hence why the datastep would not work and why I broke it up - making it inefficient, though not so much of an issue on datasets of less than 100k rows - before finally getting round to putting it back together here!

Thanks also for the tip on the Datastep Manual -- can you suggest a good one, preferably that would be available from the UK as otherwise I'd be paying express US postage from fatbrain.com or somesuch, which frankly spoils the party,

Best Regards.

Date: Thu, 25 Sep 2003 14:01:50 -0700 From: m n <iced_phoenix@YAHOO.COM> Subject: Re: Data Step Question

ben.powell@CLA.CO.UK (Ben Powell) wrote in message news:<200309251339.h8PDd3v01281@listserv.cc.uga.edu>... > I'm not clear on what can be done in any single datastep and would > appreciate any pointers. > > Say for example I want to copy a dataset to my work lib, rename a > variable and keep that and one other variable. Because in the past I > have found that sometimes a step won't action unless there is a run > command after it I would break this job into 3 seperate data steps, > which is quite repetative. Is there a rule of thumb for when a new > datastep is needed and how many steps can be included in a datastep? > > e.g. > > data a; > set lib.a;run; > data a (rename=(var1=var2)); > set a;run; > data a; > set a; > keep var2 var3; > run;

Hi Ben,

It seems that there's a misunderstanding of the data step at work here. Reading any good SAS Data Step programming manual will help; in the mean time, perhaps it will help to note a couple properties of the data step:

* Each data step is a separate program; it is compiled and executed separately from every other step in your SAS job. Thus, once you start thinking of the data step as a separate program in its own right, you can start to exploit its tremendous power.

* Each data step is an implicit loop. A SAS Data Step manual will explain A LOT about this fundamental concept.

In the meantime, other posters solved your problem using data set options. To give you some further insight, the same thing could be accomplished with:

data a; set lib.a (rename (var1=var2)); keep var2 var3; run;

Picture the above data step as a loop, each iteration reading in one observation from lib.a (via the set statement) and simultaneously renaming var1 to var2, then outputting only var2 and var3. You can do a tremendous amount of other processing in this implicit loop; each iteration of the loop will carry out the processing on just the current observation.

I hope that helps, Matt