Rules for Installing UDFs on the ML Engine | Teradata Vantage - Rules for Installing UDFs - Teradata Vantage

Machine Learning User Guide

Product
Teradata Vantage
Release Number
9.01
1.3
Published
August 2020
Language
English (United States)
Last Update
2020-10-02
dita:mapPath
dci1595445931473.ditamap
dita:ditavalPath
dqp1599597541027.ditaval
dita:id
B700-4004
lifecycle
previous
Product Category
Teradata Vantageā„¢
  • The user performing the installation must be authorized. A user is authorized if they own the schema or the schema is public.
  • The UDF (or a file with the same name) does not exist on ML Engine.
  • The UDF name is different from the destination schema name.
  • The UDF file size is 238 MB or less.
  • The archive file name for the UDF is the same as the main function class in the archive that must be executed.
  • The SQL identifier for the UDF is registered through a sysuif.install_file call before invoking the PM.INSTALL_AFUNCTION call.
  • The SQL identifier name is not one of the following, which are reserved for Advanced SQL Engine JAR files:
    • javFnc
    • terajdbc4
    • tdgssjava
    • tdgssconfig
  • The SQL identifier is not the name of the destination ML Engine schema.
  • The length of the SQL identifier does not exceed 30 characters.
  • The SQL identifier follows the rules for naming ML Engine analytic functions, without enclosing the names in double quotation marks. It must not begin with a digit, and must include only the following characters:
    • Uppercase or lowercase characters (A to Z or a to z).
    • Digits (0 through 9).
    • Special characters dollar sign ($) and underscore (_).
  • For correct UDF installation, the command sysuif.install_file must be called first. The recommended way to do this is to use the Teradata Studio client.

A failed installation does not make changes on Advanced SQL Engine and ML Engine.

Rerunning a successful pm.install_afunction always fails.