Vantage Analytics Library Functions | Analytics Library - Welcome to Vantage Analytics Library - Vantage Analytics Library

Vantage Analytics Library User Guide

Deployment
VantageCloud
VantageCore
Edition
Enterprise
IntelliFlex
Lake
VMware
Product
Vantage Analytics Library
Release Number
2.2.0
Published
March 2023
Language
English (United States)
Last Update
2024-01-02
dita:mapPath
ibw1595473364329.ditamap
dita:ditavalPath
iup1603985291876.ditaval
dita:id
zyl1473786378775
Product Category
Teradata Vantage

Using the Vantage Analytics Library User Guide

Vantage Analytics Library provides the Data Scientist with over 50 advanced analytic functions built directly in the Analytics Database.
This guide is intended for use by the following technical personnel who are familiar with Teradata SQL and Stored Procedures, such as:
  • Data Analysts
  • Data Scientists

Why Would I Use this Content?

To learn the functions necessary to perform in-database predictive modeling, evaluation, and scoring.

How Do I Use this Content?

As a reference for the functions that support the entire data science process— exploratory data analysis, data preparation and feature engineering, hypothesis testing, statistical and machine learning model building, and scoring.

How Do I Get Started?

To install the Analytics Library, run the installation scripts supplied with the product using the instructions in the readme.txt file on the installation media.

Contact Teradata for assistance with installing Vantage Analytics Library.

Vantage Analytics Library Functions

To call a Vantage Analytics Library function, use Teradata Studio™ or Basic Teradata Query (BTEQ). The controlling program for these analytic functions is the custom Teradata External Stored Procedure td_analyze. Some functions also use custom Teradata Table Operators. To use the Vantage Analytics Library functions, these user-defined functions (UDFs) must be installed in the database using scripts supplied with the product.

To run the Statistical Tests described in this guide, the Statistical Test Metadata tables must be loaded into a database on the system to be analyzed. You can use the supplied scripts to accomplish this. (Each Statistical Test function has a statsdatabase parameter for specifying the database in which these tables are installed.)

You can use the Analytics Library with any SQL processing tool that Teradata supports.

  • You must qualify td_analyze calls with the database where it is installed, for example: val_user.td_analyze.
  • By default, before an analytic function creates an output table, it drops the existing table. To prevent dropping existing output tables, specify overwrite=false.
  • You can copy and use the examples in this document on a system where the UDF is installed and the tutorial environment is set up. The tutorial environment includes the val_user default user name, val_source database with tutorial tables installed, and val_results database.
  • Starting with SQL Engine Release 16.20 Update 2, during the initialization of Vantage on newer platforms, the Always Fallback property is selected. Therefore, any fallback parameter set in the analytic functions is ignored. However, for older systems or for systems upgraded from them, the fallback parameter for analytic functions may be used effectively.