Date:         Wed, 21 Nov 2007 06:44:32 -0500
Reply-To:     Peter Flom <peterflomconsulting@mindspring.com>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Peter Flom <peterflomconsulting@MINDSPRING.COM>
Subject:      Re: What simple tutorials do you need for Base SAS?
Comments: To: RolandRB <rolandberry@HOTMAIL.COM>
Content-Type: text/plain; charset=UTF-8

Since you ask..... here are some basics that would be nice to have in one spot, although there are already papers and such on many of them, and a lot in the SAS-L archives

1.  When going from a long to a wide data set, when to use the data step and arrays and first. and last., and when to use PROC TRANSPOSE.

2.  Related to above, some complex cases of going from one format to the other.  What if you have a LOT of variables and a fair number of repeats of data?  To be specific, I currently have a data set with 1900 variables, each repeated from 2 to 8 times, for about 100 subjects.  Each repeat has an associated date, and eventually I will want to do a variety of analyses with the data, possibly including survival analysis (which would necessitate forming durations), and multilevel modeling (PROC MIXED stuff).

3.  Ways of organizing particular programs.  When programs get reasonably long. with mixes of DATA steps and PROCs, and different data sets being used for different things, how to document these and organize them so that it's coherent and relatively easy to use?

4.  How to organize sets of programs.  Is it better to have lots of little programs, each of which does a specific thing? Or a few big ones, each of which does many things?  How about all the DATA step stuff in one program, and the PROC stuff in several smaller programs?

All stuff that puzzles this data analyst who needs to do a little programing to make his analysis work :-)

Peter