Would be nice to see a new thread with new questions!
Normally data is distributed in spool on the joining columns hash value. The rule of thumb is that for joining the data has to be on the same AMP. If a PI is joined to a non-PI column, then the non-PI column data is redistributed in spool based on the joining column to make sure that the joining data is available on same AMP.
You can look for further details in Teradata design documentation.
↧