Archives of the TeradataForum

Message Posted: Mon, 27 Sep 2005 @ 00:11:27 GMT

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

Teradata doesn't use B-Tree indexes so the ordering of the columns is irrelevant. It creates a subtable that maps the secondary index values to the actual row id's. The downside is that you either use all, or nothing from a multicolumn index.

You need to define a primary index in the create table statement. The primary index is used to distribute the data across the amps and is mandatory. If you don't define a PI when you create the table it will just use the first column, which may be very subobtimal.

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