If the tables are in the same database, just perform the cross-join there. Your DB components can handle ANY legitimate SQL.
If they are not in the same DB, you can achieve this by using two tMap components in the format shown below....
Dataset B Dataset C
| |
Dataset A ------> tMap_1--------->tMap_2------------>cross-joined content
Your Dataset A is fed into the tMap_1 as the Main flow. The DatasetB is a lookup which is not joined. Output all columns from that. Then connect another tMap with the output of tMap_1. Add a lookup with Dataset C. Do not join the lookup. Output all of the columns.
Every row that enters via the Main flow of tMap_1 will be multiplied by every row in Dataset B. Likewise, every row that enters tMap_2 will be multiplied by Dataset C.
