Dynamic Data Flow - How to create completely dynamic data flow?
Starting from 1.4 SR-3 the data flow can be completely dynamic - the need to have one common column across different iteration sets is no longer required. To have completely dynamic components, you have to remove all statically defined columns (both in the components and related connection managers). However there are components, which require at least one column to pass successfully validation. We have implemented a workaround, which is based on thunk (not used) column. If you encounter a component requiring at least one column, define a column named THUNK_COLUMN. This column will be maintained and ignored at runtime by the Data Flow Task Plus.