17.05 - INSERT Process - Advanced SQL Engine - Teradata Database

Teradata Vantage™ - SQL Data Manipulation Language

Advanced SQL Engine
Teradata Database
Release Number
January 2021
English (United States)
Last Update
An INSERT process performs the following actions:
  1. Sets a WRITE lock on the rowkey, partition, or table, as appropriate.
  2. Performs the entire INSERT operation as an all-or-nothing operation in which every row is inserted successfully or no rows are inserted.

    This is to prevent a partial insert from occurring.

Session Mode Unsuccessful INSERT Result
ANSI Rolls back the erring request only.
Teradata Rolls back the entire containing transaction

The rules for rolling back multistatement INSERT requests for statement independence frequently enable a more relaxed handling of INSERT errors within a transaction or multistatement request. For information about failed INSERT operations in situations that involve statement independence, see Multistatement and Iterated INSERT Requests.

The INSERT operation takes more processing time on a table defined with FALLBACK or a secondary, join, or hash index, because the FALLBACK copy of the table or index also must be changed.