@rxhmx wrote: Dear all,    I have a project in which I have to check automatically the evolution of some variables each month  For instance, I have to chech the frequency of each modality for a variable in the table n-1 compared to the table n  These variables are dummies and can have until 100 modalities.  The databases are really big and I'm looking for a way to extract from each table the variables with categorical data (there is more than 300 variables) and merge them two by two  Sure this is possible, I think you would probably need to perform PROC FREQ on each data table, then a macro of some sort to loop through the different variables and compare the results of the PROC FREQ across the two tables. Although first I would investigate if PROC COMPARE does enough to meet your needs.   You phrase this problem in terms of "dummies", but if I am understanding you properly, there is no need for dummy variables at all. ... View more