All Forums Database
soumyajit 2 posts Joined 12/12
23 Jul 2013
Difference between ANSI and TERA modes with respect to record insertion

I am working on an application that takes records from Hadoop and inserts them into Teradata via sqoop(JDBC).
I am using TERA mode for the connection and the target table is SET table. I am getting duplicate row issue for some datasets.
As far as I know TERA mode is supposed to ignore duplicate records while doing inserts. Can somebody please confirm the behavior?

CarlosAL 512 posts Joined 04/08
23 Jul 2013

It has nothing to do with Transaction mode (TERA/ANSI), It's all about the table definition (SET/MULTISET)

tomnolan 594 posts Joined 01/08
24 Jul 2013

soumyajit, you were misinformed. The Teradata Database never "ignores" duplicate records when doing inserts, no matter which transaction mode (ANSI vs TERA) you are using.
As Carlos said, the table attribute (SET vs MULTISET) governs whether duplicate rows are permitted in a table. With a MULTISET table, you are permitted to insert duplicate rows using either ANSI or TERA mode.
In contrast, with a SET table, if you try to insert duplicate rows, you will get Teradata Database error 2802 "Duplicate row error in <table>" using either ANSI or TERA mode. In other words, the transaction mode does not change this behavior.

You must sign in to leave a comment.