Archives of the TeradataForum

Message Posted: Thu, 12 Aug 2010 @ 18:23:51 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

If you INSERT/SELECT to a NoPI table there is no data redistribution for the final merge. The rows will all be inserted locally, on the same AMP where they were located in the source table or spool file.

On the other hand, if you are loading the NoPI table from a client, the data will be sent round-robin to the AMPs and distributed equally.

You say this is a staging table. You should define the PI based on what you are going to do next with the data. If the next step is to join or merge with a target table, you should normally make the staging PI match the target. NoPI may be appropriate if you are loading from an external source and then must do some transformation or aggregation before you can apply to the target.

	<Prev	Next>		<<First	<Prev	Next>	Last>>