This chain shows two examples of preparing data for Apriori, the well-known association rule learning algorithm. One example applies the Apriori implementation which comes with MiningMart; on the assumed data format, this requires no preparation. However, if other Apriori implementations are going to be used, they probably require some preparation which ensures that each data row corresponds to a single transaction. Thus the other example in this chain shows how this can be done. It adds binary indicators for the presence of a product to each row in the input data. For example, in a row that indicates that product number 3 is sold in the transaction 16, a new attribute for product number 3 takes the value 1 while new attributes for all other products take the value 0. This allows to group rows by transaction, taking the maximum of the new product indicators so that all products sold in a transaction get value 1, while all products not sold in that transaction get value 0. This is the desired input format for an Apriori implementation like the one in YALE.