Aggregation

The transformation of record-level data to a higher level of analysis. An exceedingly common task, but one that requires a thorough cheat-sheet documenting the 'gotchas' and tricks.

Contents

  1. Aggregation
    1. Implicit Aggregation

Implicit Aggregation

Several procedures compute intermediary variables to produce a report. These intermediary variables can be stored and used.

To count observations by catgegory, use proc freq. 

proc freq data=LIBREF.TABLE noprint;
  tables=VAR / out=LIBREF.OUT;
run;

Many higher-level statistics can be retrieved through proc means. For example, to compute VARMEAN and VARMAX from VAR1 for each category of VAR2. 

proc means data=LIBREF.TABLE noprint;
  var VAR1;
  class VAR2;
  ways 1;
  output out=LIBREF.OUT(drop=_FREQ_ _TYPE_) mean=VARMEAN max=VARMAX;
run;

CategoryRicottone