CC09
Change Shape or Lose a Few: Data Set Reduction for Analysis
Contributed
Philip J. d'Almada
Consultant
Abstract:
Given a multifunctional or event-based stream of variables, the task at hand for a particular analysis is to remove events or data values that are not needed. SAS software is invoked in either one of two methods whereby a new analytic data set is created t
hat differs from the parent data set by either having an increased number of records and a single representative variable for each source variable stream, or has no change in the number of records but fewer variables by effectively discarding useless value
s. The OUTPUT feature is invoked for the first solution. For the second solution, the ARRAY statement provides the key to data reduction in a technique involving two DATA steps and a CALL routine that invokes the SAS macroprocessing facility.
Biography:
Mr. d'Almada began using SAS in 1976 but became a full-fledged SAS user 12 years later. He presently uses SAS software to support research in public health, specifically, paediatric HIV/AIDS research. He has presented at previous SESUG conferences and enco
urages any SAS user to do the same.