If the variable names are different, then one can use the data set rename option.

For example, q1(rename=(q1=q))

set q1(in=a rename=(q1=q) ) q2(in=b rename=(q2=q)) q3(in=c rename= (q3=q));

will do the trick. HTH

dc353@hotmail.com wrote: > On Jan 13, 11:42�am, shilin...@yahoo.com wrote: > > It is not a �Cartesian product but a set of all data sets. > > > > Here is an example. > > HTH > > > > data all; > > � set q1(in=a ) q2(in=b) q3(in=c); > > � group=(a=1)*1 + (b=1)*2 + (c=1)*3; > > run; > > > > On Jan 13, 9:57�am, "dc...@hotmail.com" <dc...@hotmail.com> wrote: > > > > > > > > > I have the following data: > > > > > � q1 > > > � 5 > > > � 8 > > > � 4 > > > > > � q2 > > > � 1 > > > 10 > > > > > � q3 > > > � 1 > > > � 2 > > > � 3 > > > > > ... > > > > > each data set represents a question and each row represents an answer > > > to that question. �I have N questions each with M(n) answers. �I > > > wondering how to create a Cartesian product between all N questions. > > > In the case where N = 3 from the above data the answer would like: > > > > > 5 1 1 > > > 5 1 2 > > > 5 1 3 > > > 5 10 1 > > > 5 10 2 > > > 5 10 3 > > > 8 1 1 > > > 8 1 2 > > > 8 1 3 > > > 8 10 1 > > > 8 10 2 > > > 8 10 3 > > > 4 1 1 > > > 4 1 2 > > > 4 1 3 > > > 4 10 1 > > > 4 10 2 > > > 4 10 3 > > > > > I was wondering if I create the dataset as: > > > group � answer > > > 1 � � � � � � 5 > > > 1 � � � � � � 8 > > > 1 � � � � � � 4 > > > 2 � � � � � � 1 > > > 2 � � � � � � 10 > > > 3 � � � � � � 1 > > > 3 � � � � � � 2 > > > 3 � � � � � � 3 > > > > > Is there a way to do a Cartesian product by group?- Hide quoted text - > > > > - Show quoted text - > > Hi, > > Your code produces the following output for ALL: > > Q1 Q2 Q3 group > 1 5 . . 1 > 2 8 . . 1 > 3 4 . . 1 > 4 . 1 . 2 > 5 . 10 . 2 > 6 . . 1 3 > 7 . . 2 3 > 8 . . 3 3 > > However the following code produces the correct output: > > proc sql; > create table ALL as > select * > from q1,q2,q3; > quit; > > I believe this is the basic form for a cartesian product. My question > has to do with creating the input dataset as 1 set as opposed to 1 > dataset for each group and creating the cartesian product without > having to know how many groups are in the combined dataset.