X hits on this document

404 views

0 shares

1 downloads

0 comments

131 / 131

TPT: use Teradata Parallel Transporter. When using TTU KMs, you should also take into account the following KM parameters:

REPORT_NB_ROWS: This option allows you to report the number of lines processed by the utility in a Warning step of the integration interface.

SESSIONS: Number of FastLoad sessions

MAX_ALLOWED_ERRORS: Maximum number of tolerated errors. This corresponds to the ERRLIMIT command in FastLoad/MultiLoad/TPump and to the ErrorLimit attribute for TPT.

MULTILOAD_TPUMP_TYPE: Operation performed by the MultiLoad or TPump utility. Valid values are INSERT, UPSERT and DELETE. For UPSERT and DELETE an update key is required in the interface.

For details and appropriate choice of utility and load operator, refer to the Teradata documentation.

Support for Named Pipes

When using TTU KMs to move data between a SQL source and Teradata, it is possible to increase the performances by using Named Pipes instead of files between the unload/load processes. Named Pipes can be activated by setting the NP_USE_NAMED_PIPE option to YES. The following options should also be taken into account for using Named Pipes:

NP_EXEC_ON_WINDOWS: Set this option to YES if the run-time agent runs on a windows platform.

NP_ACCESS_MODULE: Access module used for Named Pipes. This access module is platform dependant. Refer to the knowledge module option help for a list of values.

NP_TTU_STARTUP_TIME: This number of seconds for the TTU to be able to receive data through the pipe. This is the delay between the moment the KM starts the TTU and the moment the KM starts to push data into the named pipe. This delay is dependant on the machine workload.

Optimized Management of Temporary Tables

Creating and dropping Data Integrator temporary staging tables can be a resource consuming process on a Teradata engine. The ODI_DDL KM option provides a mean to control these DDL operations. It takes the following values:

DROP_CREATE: Always drop and recreate all temporary tables for every execution (default behavior).

CREATE_DELETE_ALL: Create temporary tables when needed (usually for the first execution only) and use DELETE ALL to drop the temporary table content. Temporary table are reused for subsequent executions.

DELETE_ALL: Do not create temporary tables. Only submit DELETE ALL for all temporary tables. NONE: Do not issue any DDL on temporary tables. Temporary tables should be handled separately.

Knowledge Modules Reference Guide

131/131

Document info
Document views404
Page views407
Page last viewedFri Dec 09 01:30:51 UTC 2016
Pages131
Paragraphs3942
Words42855

Comments