I have been looking at the Join and Merge tools recently, and have a delayed response. There are two main differences in the Join and Merge tools in Phoenix.
The first difference is the order in the output. The Join tool internally resorts the data for processing, and therefore also for the resulting worksheet, by the columns that are mapped as Sort variables. The Sort precedence is the order in which the Sort columns are mapped. Resorting data is standard in SQL joins, which is that concept that Phoenix uses. The Merge tool however merges the data by the order of the data in the first worksheet, so the Merge result keeps the order of the first worksheet, followed by the order of the data from the second worksheet that has not already been merged in. (Note that, because Merge appends the unmatched data from the second worksheet, Merge is similar to an Outer Join, not an Inner Join, in that the result has all data from both worksheets.)
The second difference occurs only in the many-to-many case for matched values, i.e., the case where, for a specific set of values to match on, both datasets have multiple rows that have those values, so a user might never run into this case. The Join tool will use a Cartesian join in this case, sometimes also called a cross join. As an example, if the first worksheet has three rows with the match values, and the second worksheet has four rows with those match values, the result will have twelve rows for the match values, which contain all combinations of the values from the two worksheets. Again, this is standard for SQL joins. The Merge tool however will have the larger number of rows from the two worksheets, in this example, four rows. The Merge tool will combine the rows sequentially until it runs out of matched data in one of the worksheets. Then it will either leave a blank cell for the worksheet that has run out of matched data, or will carry down the last available value from that worksheet, and the option “Carry Along Data for Like Sort Levels” controls which method is used.