17.10 - Usage Notes - Advanced SQL Engine - Teradata Database

Teradata Vantageā„¢ - DATASET Data Type

Advanced SQL Engine
Teradata Database
Release Number
July 2021
English (United States)
Last Update

DATASET_TABLE resides in TD_SYSFNLIB. The query result ID column is a number or character type, excluding CLOB, and must be unique.

The result is subject to a maximum row size, and the query cannot exceed the maximum allowable query size.

column_expression_literal requires mapping the RowExpr columns to the function output table columns. The output table column data type is any non-LOB predefined Teradata type specified in the column_expression_literal. Each column in column_expression_literal is defined by a JSON object conforming to normal column or ordinal column.

Example of a normal column:
{ "dotnotation" : "<path to source data for output column>",
	  "type" : "<data type of output column>",
	  "fromRoot" : true (if search is to be done from the root, instead of the row expression result) }
Example of an ordinal column:
{ "ordinal" : true }

For CSV files, default assignment for the source column to target table columns is based on position. The first CSV column is assigned to the first table column, the second CSV column is assigned to the second table column, and so on. The CSV data does not have to align with the target table being loaded. Since CSV data is shredded into a temporary table, use the optional method to INSERT or SELECT between tables with columns not aligning perfectly.