if you use a field from a table that was never used in the report before you may now be 'enforcing' a join which just changed the size of your data set and may have turned a small data set into a huge data set (especially if you did not join the table giving you a cartesion result).

crystal does not enforce joins unless you explicitly set the enforement in the join set up or if you use fields from both tables in the join.

Add a count (not distinct count) of any field you already have in the report canvas (not just in your data set).

Place that in the report header. Note the result.

now drag the commission field into the report header and watch to see if your count changes.

in the sub report look at the Database screen and then the joins tab.

Look to see if all the tables are joined. If not I am guessing the table that the commision field is in is not joined at all.

if that is the case, when you added it to the report it enforced the lack of a join and you get a cartesion result which is every permuation of row to row connection (a table with 10 rows and another table of ten rows will give you a dataset of 100 rows, not 10).

just to test it again get it back to the OK performance and drag a different field from the same table as the commision into the sub report and see if you get the same issue.