SAS PROGRAMMING
                                          HANDOUT #5
                                       DATA SET MERGES


     There are several ways to combine and/or subset data sets.

     Suppose  AA  and  BB  are data sets, with number of observations na, nb.


DATA CC; SET AA;         This just gives a new name to the data set AA.
  DROP X1 X2;            And drops the X1 and X2 variables.
  WHERE X=2;             These two statements select observations with X=2
  IF X=2;

DATA DD; SET AA;
  IF 1<X<5;              The data set DD is a subset of the data set AA.
  IF 3<X<5 THEN Y=1;     This creates a new variable Y with value 1 when 3<X<5.
  KEEP X3 X4;            And keep the X3 and X4 variables

DATA EE; SET AA BB(RENAME=(old=new));
                         The data set EE has  na+nb  observations.
                         If X is a variable in AA, but not in BB then X is missing
                            for those observations coming from BB.
                         The old variable in BB is renamed new in the new data set.


DATA FF; MERGE AA BB;    The data set FF has  max(na,nb) observations.
                         The observations are matched in order.
                         If X is a variable in both AA and BB, it takes on the value
                             from the data set BB.

PROC SORT DATA=AA; BY X1;
PROC SORT DATA=BB; BY X1;
DATA GG; MERGE AA BB; BY X1;

                         The observations from AA and BB are matched using the X1 variable.
                         If X is a variable in both AA and BB, it takes on the value
                              from the data set BB.
                         If there are more than one observation corresponding to an X1 value
                              then there is a many to one matching.

PROC SORT DATA=AA; BY X1;
DATA HH; SET AA; BY X1;

                         The data set HH has a new variable FIRST.X1 that is equal to 1
                               for each observation that is the first for an X1 value.
                               It also has a variable called LAST.X1

PROC SORT DATA=AA; BY X1 X2;
DATA PP; SET AA; BY X1 X2;

                         The data set PP has new variables FIRST.X1, FIRST.X2, LAST.X1 and LAST.X2

PROC SORT DATA=AA; BY X1;
DATA JJ; SET AA; BY X1;
 IF FIRST.X1=1 THEN NUMB=0;
 NUMB+1;

                         The data set JJ has a variable NUMB that keeps track of the number of
                              observations corresponding to each X1 value.

DATA KK;
 DO I=1 TO 10;
  Y=4+2*RANNOR(0);
 OUTPUT; END;

                         The data set KK has 10 observations from a normal mean 4, sd=2;