Restrictions and Limitations - MultiLoad

Teradata MultiLoad Reference

Product
MultiLoad
Release Number
16.10
Published
May 2017
Language
English (United States)
Last Update
2018-07-11
dita:mapPath
cgb1488824663145.ditamap
dita:ditavalPath
Audience_PDF_product_tpt_userguide_include.ditaval
dita:id
B035-2409
lifecycle
previous
Product Category
Teradata Tools and Utilities

The following table describes the Teradata MultiLoad restrictions and limitations on operational features and functions.

Feature Restrictions  
Operational Feature/Function Restriction/Limitation
Aggregate operators, exponential operators, arithmetic functions Not allowed
Concatenation of data files Not allowed
Concurrent load utility tasks The maximum number of concurrent Teradata MultiLoad tasks that can run is variable; the limit can be controlled by the system administrator. MaxLoadTasks may be overridden if TASM is active.
For the most up-to-date information on concurrent task limits, see the description of the MaxLoadTask parameter of the DBSControl utility in Utilities (B035-1102). Additional information is available in the Teradata Dynamic Workload Manager User Guide (B035-2513).

If the Teradata MultiLoad job exceeds the recommended limits, Teradata Database returns a 2633 error message indicating that too many loads are running, and the utility retries until:

  • It can execute the task
  • It reaches the TENACITY hours time limit specified by the BEGIN MLOAD command
Data retrieval from Teradata Database with the SELECT statements Not allowed
Expressions Are evaluated from left to right, using the Teradata Database order of preference, but can be overridden by parentheses.
Foreign key references Not allowed for a Teradata MultiLoad target table.

Attempting a Teradata MultiLoad task or any other action against a target table defined with a foreign key constraint produces an error condition.

Hash indexes Not supported
Hexadecimal form Teradata MultiLoad does not accept and will not display object names specified in internal Teradata Database hexadecimal form.
Join index restrictions Teradata MultiLoad does not support or maintain Join Indexes.
Journaling options There are no restrictions on journaling options. Teradata MultiLoad recognizes target tables with all types of currently supported journaling.
Maximum number of columns 2048
Maximum row size The maximum row size for a Teradata MultiLoad job, data plus indicators, is approximately 64,000 bytes. This limit is a function of:
  • The row size limit of Teradata Database
  • The Teradata MultiLoad work tables and error tables

Teradata MultiLoad cannot accommodate a row size that is near the row size limit of Teradata Database because of the operational overhead associated with Teradata MultiLoad work tables and error tables. If the Teradata MultiLoad job exceeds this limit, Teradata Database returns Error Message 3577 (row size or sort key size overflow), and terminates the job.

Maximum statement text length 1 MB
No primary index table Not allowed
Normalized Tables Not allowed
Unique secondary indexes Not allowed for a Teradata MultiLoad target table.

Target tables for a Teradata MultiLoad task may contain NUSIs, but not unique secondary indexes.

Tables with referential integrity Not Allowed.
Tables with defined triggers Not Allowed.

PPI Table Considerations

Load utilities, including Teradata MultiLoad, are supported on PPI tables with the following restrictions and advisories:

  • Teradata MultiLoad does not support tables with Unique Secondary Indexes (USI). Many PPI tables have USIs.
  • Teradata MultiLoad IMPORT tasks require all values of the primary index column set and all values of the partitioning column set for deletes and updates.
  • Teradata MultiLoad IMPORT tasks do not support updates of the partitioning column set.
  • Teradata MultiLoad IMPORT tasks do not support primary index updates.
  • Teradata MultiLoad does not support hash and join indexes.

NoPI Table Considerations

A NoPI Table is a table that has no primary index. These tables can be used as staging tables where data is always appended to the table, making population of the table generally faster than that of a traditional table containing a primary index.

See the No Primary Index discussion in SQL Data Definition Language (B035-1144).