You can do that pretty easily in PROC TABULATE if you don't mind having multiple variables there (ie, not just VAR_1 with '1 3 0' but three columns).

-Joe

On Tue, May 24, 2011 at 11:42 AM, Warren Schlechte < Warren.Schlechte@tpwd.state.tx.us> wrote:

> Hopefully, the formatting of this message will make it easier to read. > ------------------------------------------------------------------------ > --------------------------------- > > I think this should be simple and not require me to create a new > dataset, but I cannot see how to do this unless I first create a dataset > that gives me the sums for each variable and value. > > What I have is a dataset that looks like: > > Var_1 Var_2 Var_3 Ind_Var > A C B 0 > B A B 0 > A C A 1 > B B A 0 > C A C 1 > A B C 1 > B C C 0 > > What I want is a table where the cell entries are the sum of the > "cross-products" of the indicator variable and the variable values, > like: > > Var_1 Var_2 > Var_3 > A B C A B C A B > C > Ind_Var > 1 2 0 1 1 1 1 1 > 0 2 > 0 1 3 0 1 1 2 1 > 2 1 > > > Hopefully, this is a simple procedure call, just one I don't use very > often. > > If it requires programming, I'm interested in seeing how others would do > this, but don't feel compelled to respond as I can get something > together. > > Thanks, > > Warren Schlechte >