Make the dataset have one observation for every possible interaction of a list of variables.
fillin(varlist)
a variable list in "var1 var2 var3 x*" format where "*" matches zero or more of any character and "?" matches one of any character (or a varlist in formula format, ~var1+var2+var3+x1+x2+...). On exit, the data set will contain one observation for every possible interaction of variables with missing values filled in where appropriate.
returns NULL, invisibly