Thank you for your answer!   Well, all the A variables are character, all the B are numeric, all the C values have a date format (not in this example, but in my real dataset).  Actually, A variables contain first names and last names (of people receiving subsidies), but not all of the A variables (coming from different data providers but merged in a big dataset) contain these names as I have indicated it in the example. These cells are fully empty - I designated them as missing values. These missing values cause difficulty for me. In fact, you highlighted that treating these missing values was the key issue here, because they could 'cheat' or mislead me in comparing the values.    My goal is to compare every cell that contain a value (so not empty) by row (concerning my real dataset: my purpose is data correction and imputation).    Is there a quicker or simplier solution for this than comparing the variables one by one using e.g. an 'ID' variable indicating which of the A1-4 variable contain a value (e.g. 0101 (so A2 and A4), 1100 (so A1 and A2))?    ... View more