MINIME:  Reset variable lengths to minimize the size of a SAS dataset

Programmer:  Dan Blanchette ()
Center for Entrepreneurship and Innovation
Duke University's Fuqua School of Business
Durham, NC  USA

Date:  22May2003
Last updated:  09Mar2009

---

Reset the variable lengths to minimize the size of a SAS dataset

%minime( variables not to minimize , variables to minimize );

The MINIME SAS macro resets variable lengths of the dataset most recently created in the WORK library to the safest minimum for all operating systems.  This should reduce the size of the SAS datafile, but it could increase it, perhaps appropriately so.  This is like the Stata command compress.

NOTE:  The parameters have to be specified in the order of this list.

Parameters	Explanation
(1st) variables not to minimize	The first string inside the parentheses can be a list of one or more variables not to minimize.  In general, it is not good to minimize long integer variables like ID variables in case in a future merge the dataset being merged in has longer ID variables than the ones in the current dataset being minimized.  If no variables are listed not to minimize, then the default is to minimize all variables.
(2nd) variables to minimize	The second string inside the parentheses after the first comma can be a list of one or more variables to minimize.  If no variables are listed to minimize, then the default is to minimize all variables.

Using the MINIME SAS macro requires that you understand how to use the %include statement and that you know how to call a SAS macro.

  %include "LOCATION AND NAME OF A FILE THAT CONTAINS SAS CODE";

For example, if you have copied this file to "C:\SASmacros", then you tell SAS about this macro by adding the following line to your SAS program:

  %include "C:\SASmacros\minime.sas";

The include statement makes SAS aware of the MINIME SAS macro which is in the file minime.sas.  You call the macro by adding this to your code:

  %minime;

 %include "C:\SASmacros\minime.sas"; ** Make SAS aware of the MINIME macro.**;

 data work.new;
  set work.old;
 run;

 ** now call the MINIME SAS macro to minimize all the variables in the dataset **;

 %minime;

 ******************************************************************
  -- To see the attrib statement MINIME used to reset the variable 
      lengths, use the PRINTME SAS macro after MINIME has run:
 ******************************************************************;

 %printme;

 proc contents data= work.new;
 run;


 ** or all character variables **;
 data work.new;
  set work.old;
 run;

 %minime( , _character_ );  ** notice that the first field is empty and
                             *  that a comma separates the two fields **;

 proc contents data= work.new;
 run;


 ** or only the numeric variables except for you ID variable: **;
 data work.new;
  set work.old;
 run;

 %minime( companyid , _numeric_ ); 

 proc contents data= work.new;
 run;

 ** or all variables except for you ID variable: **;
 data work.new;
  set work.old;
 run;

 %minime( companyid ); 

 proc contents data= work.new;
 run;