PurpuraInk Consulting- Chicago, IL & San Juan, PR

Advanced Topics on SAS Data Step Programing: What are
your options?

last updated: 2FEB04

System options

Before you start, you might be interested in knowing what options are
selected as default in your SAS session. If you run:


PROC options; run;


you will see in the LOG window a complete listing of what are the options
you are using in your SAS session. You do not have to worry about many of
the things you will see displayed, so don’t panic.

Page set up options.

This code… options nodate pageno=1 nocenter linesize=120 pagesize=65 FORMCHAR="|----|+|---+=";	Results in… Page numbering starting with one width and height of the paper to larger area supress SAS dates supress centering of the output on the output page set a character formatting for lines such that it is easier to copy and paste into other programs.
This code… TITLE1 "Here is my title %SYSFUNC(DATE(), EURDFDE9.) %SYSFUNC(TIME(), TIMEAMPM8.)";	Ouputs as: Here is my title 02OCT2003 5:07 PM 17:04 Thursday, October 2, 2003 1

 

 
 
 

DATA step options


 

The DATA step has two types of options: one which modifies the individual
datasets, and one which modifies the DATA step itself. We will refer to
those who make global changes within the DATA step as system
options.


DATA dataset1 (OPTION =list1) dataset2 (OPTION =list2) /OPTIONS=list;
…
RUN;




1. Debugging your programsSometimes it is convenient to step through the code SAS is processing. The
   Data Step Debugger is a tool that allows you to do this. To invoke it, you
   use the option DEBUG:

   
   DATA dataset1 /DEBUG; RUN;
   

   As a result a SOURCE window will appear. This window displays the data
   step code and the actual line of code being executed. The LOG window will
   accept debugger commands. Some of categories examples of debugger
   commands are:

   GO JUMP STEP	<– Program execution control
   CALCULATE DESCRIBE EXAMINE SET	<– variable manipulation
   BREAK DELETE LIST TRACE WATCH	<– debugger request manipulation
   ENTER	<– tailoring
   QUIT	<– terminate
   HELP SWAP	<– window control

2. Managing Variables [list1 or list2].Each of the following three commands can be used, and/or combined, to
   manage
   varibles in one of three basic ways.
   1. Within a single file
   2. Within Multiple files
   3. Across multiple files

   These commands are:

   KEEP= variable list	reads/saves only listed
   DROP= variable list	reads/saves all but listed
   RENAME= (oldname=newname)	substitutes names of listed

   For Example:

   libname cdir 'c:\temp\';
   DATA myfile1 (DROP = year age) myfile2 (KEEP= index origin RENAME=(ss=social) ;
   set cdir.myinput1;
   run;

   …will read in C:\temp\myinput.ssd and create two files: myfile1
   and myfile2 which will be modified by dropping or keeping variables
   Or, to drop a list of variables from BOTH myfile1 and myfile2, you would
   take advantage of the OUTPUT statement which is implied at the end of
   the data step by typing…

   
   

   DATA myfile1 myfile2;
   ...
   	DROP ss year age;
   RUN;

3. Managing datasets.

   Each of the following commands can be used, and/or combined, to
   manage your datafiles. Some of these commands can be used as individual
   file options or as global system options.

   <!–

   –>

   COMPRESS= [yes | no]	By default SAS dataset are stored as fixed record lengh. This option is optimally used when variable record lenght is used, as to reduce storage space.
   REUSE=	By default, space associated with data deleted within a dataset is not recycled, but left empty. This option tells SAS to reuse empty space resulting from deleted records.
   $POINTOBS=	Used for compressed datasets, allows to access a particular record.
   REPLACE= [yes | no]	Setting this option to NO, provides write protection for that file for the duration of the datastep.
   LABEL= ‘dataset label’	Assigns a label to be displayed by procedures. This helps is the file name is cryptic.
   SORTEDBY= var1 var2	Defines variables in the input data which are already sorted and allows for BY statement to be used in PROC or DATA steps.

4. Access control options.

   The following options control file access to SAS datasets. This means
   that you can control who reads/writes to/from your datasets.
   Remember that when passwords are used, forgetting the password means
   ‘bye-bye’ data. All of these commands use the following format.
   

   libname cdir 'c:\temp\';
   DATA myfile1 (COMMAND = options) ;
   ...
   run;

   PW=xxxxxxxx	Sets an read/write password for a dataset.
   READ=	Password only for READing a file.
   WRITE=	Password only for WRITE’ng a file.
   ALTER=	Password only if changes are made to the dataset.
   ENCRYPT= [yes | no]	Allow for encryption of dataset. Useful for protection of sensitive data.

5. Indexing

   By indexing a SAS dataset a look up table is created which allows you
   point to particular values of the indexing variable. An index can be based
   on a single variable, or on a combination of variables. If a combination
   of variables is desired, you need to give a nickname to that particular
   combination.
   

   DATA myfile1 (index = var1 ) myfile2 (index= (nickname= var1 var2));
   ...
   run;

6. Selective OUTPUT of data
   By taking control of the OUTPUT statement, you can create multiple files
   using the same SAS codes and add records to
   each file depending on certain conditions being satisfied. Note that
   this code would require for an explicit output statement for all the
   files being created.

   Here we create two files depending on the value of a variable:

   DATA myfile1 myfile2 ;
   set cdir.myinput1;
   if region='NE' then output myfile1;
   else output myfile2;
   run;

   Here we create two files: one depending on the value of a variable,
   another with all the cases:

   DATA myfile1 myfile2 ;
   set cdir.myinput1;
   if region='NE' then output myfile1;
   output myfile2;
   run;

 

 
 
 

File transfers


 

1. Reading in a SAS dataset which was created in another version of
   SAS

   To be able to read in a file which was created in another version of SAS,
   but in the SAME operating system you are intending to use this file in,
   you first need to set up a LIBRARY alias for all files for the older
   version. SAS wants to have files from one version separate from
   files of other SAS versions. In the example below we setup a directory
   where all SAS 6.12 files will be located.

   libname cdirold v612 'C:\SASDATA\data612\';
   libname cdir 'C:\SASDATA\data\';
   

2. Reading in a SAS dataset which was created in another operating
   system

   To tranfer files across operating systems, you need to save the data in
   the source system to a SAS PORTABLE file format. This format extracts all
   system specific information and leaves the data with data relevant
   information on the header of the file. This PORTABLE file is a binary file
   but smaller than its original version. Upon doing the transfer via FTP or
   the file transfer protocol of your choice, you need to read in the data
   via one of a few methods.
   Also refer to Converting
   to/form SAS files for more detailed information on this process.

   libname cdir 'C:\sasdata\newdata\';
   filename trans 'C:\sasdata\oldfile.por';
   proc cimport library=test2 infile=trans; run;
   

 

 
 
 

Macro Programming


 

Detailed introduction to Macro Programming is available at Introduction
to SAS Macro Programming .

 

 
 
 

Examples


 

• Subsetting and
  Resampling in SAS
  

 

 
 
 

References


 

• The Data Statement:
  Efficiency Techniques S. Riba, JadeTech
• You can RUN but your
  data cannot Hide: Advanced Methds to Introduce External Data into
  the SAS system

<!–

• Basic Methods to Introduce External Data into the
  SAS system

  More [and complex] Examples
  
  

  –>

   

   

  ---

  2004-2-3  VDC: WWWSTATS@uic.edu