122 Unleashing DB2 10 for Linux, UNIX, and Windows
8.1 Overview of the ingest utility
To make mission-critical decisions in the business environment of today, data
warehouses must have data that is current and available 24x7. DB2 10 features
the ingest utility, a new data movement utility that meets these demanding
requirements and can be a key component of your extract, transform, and load
(ETL) operations.
The ingest utility is a high-speed, configurable, client-side DB2 utility. With its
multi-threaded architecture, the ingest utility can pump massive amounts of data
quickly and continuously into DB2 tables from files, or pipes, with minimal effect
on concurrent user workload and data server resources.
The ingest utility differs from other data movement utilities in several key areas:
򐂰 With ingest utility, queries can run concurrently against the DB2 tables that
are being loaded with minimal effect on data availability. In other words, you
no longer need to choose between data concurrency and availability.
򐂰 The ingest utility features high fault tolerance and recoverability.
򐂰 Unlike other data movement utilities, the ingest utility supports SQL
expressions (including predicates and casting), so in addition to doing the
load stage of your ETL process, the ingest utility can also participate in the
transform stage.
򐂰 The ingest utility supports a range of DML statements, including INSERT,
UPDATE, DELETE, MERGE, and REPLACE.
򐂰 Not only is the ingest utility compatible with IBM InfoSphere® Warehouse, but
it is “database-partition-aware,” which means the utility routes inserted rows
directly to the appropriate partition (rather than going through a coordinator).
This approach contributes to its speed in a partitioned database environment.

Get Unleashing DB2 10 for Linux, UNIX, and Windows now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.