Archives of the TeradataForum

Message Posted: Wed, 20 Jun 2007 @ 19:54:02 GMT

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

The problem with this is:

          > Create databasename.table
          >    (as sel * from databasename.table name)
          >    with data;

You will most likely create a set table with a NUPI on the first column. If there are a lot of hash collisons it will slow down the copy process looking for duplicate rows.

As Fred mentioned, you should at least use a PI definition, but probably better to use the original table to get the same PI and any/all SI definitions as well as any constraints that might exist on the original table.

Just an added thought,

Michael Larkins
Certified Teradata Master
Certified Teradata SQL Instructor

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