Archives of the TeradataForum

Message Posted: Thu, 20 Dec 2007 @ 10:57:35 GMT

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

The code used by the "Phantom UDF" is real and I have used it with good results.

I understand it is unlikely to be released outside Teradata anytime soon though as it is not part of any Teradata supported offering, and I am sure you will agree a facility like this is prone to misuse. It's fine for 10/100/1000 rows here and there, but any expectations that it is the right solution for high volume cross-system joins need to be tempered by reality.

If you want to do this stuff and can access your foreign DBMS using C/C++ code, I suggest you read up on table functions and connect the dots just as the tdat function does - just remember that you will be pulling data from the foreign system into Teradata TPA node memory using TPA node cycles and potentially affecting overall TD performance. Extensive testing is essential and NOT on the production system.

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