Archives of the TeradataForum

Message Posted: Fri, 02 Nov 2012 @ 17:31:42 GMT

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

You are correct. NUSI index entries are stored on the same AMP as the row they point to, while USI index entries are hash distributed on the index columns.

A consequence is that any access via NUSI is inherently all-AMP, while USI retrieval of a single row is a two-AMP operation.

Note that it is also possible to define a Hash Index or Join Index that effectively functions as a "hashed NUSI".

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