Archives of the TeradataForum

Message Posted: Thu, 17 Jun 2004 @ 19:57:38 GMT

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

<-- Anonymously Posted: Thursday, June 17, 2004 15:53 -->

Our project team is dead set on using RI on one of our databases. Current process using a ETL tool involves bringing table dumps out of Teradata, creating a hash file, and then doing another insert back to Teradata. I have convinced them that this is a very poor design choice (actually made by Oracle ETL specialists).

I am now faced with several techniques to achieve the end result. Please help me choose the best one!

I need counts on # rows inserted, # rows deleted and # rows with RI failure.

1) turn off RI,create table, join insert, left join insert,drop table, rename table

2) write a stored procedure

3) Multiload or T-Pump job

4) ?

I am leaning toward a Multiload job, but am not sure how it will handle the RI failures. Also, will it give me the counts I am looking for?

Also, currently these are small tables with few updates/inserts but will continue to grow over time.

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