1.1 - 8.10 - FrequentPaths Input - Teradata Vantage

Teradata Vantage™ - Machine Learning Engine Analytic Function Reference

Teradata Vantage
Release Number
October 2019
Content Type
Programming Reference
Publication ID
English (United States)
Table Description
InputTable Contains sequence data to process, in either of these formats:
Format Description
Sequence/path Each row contains a string of format '[item[,...]]'. Outer brackets belong to string (for example, '[A, B, C, D]').

To output strings in this format, use Introduction to nPath function with its Accumulate syntax element.

Item Each row represents one item in a sequence.

With this format, specify either ItemDefinitionTable or ItemColumn syntax element.

ItemDefinitionTable [Optional] Required if you omit both ItemColumn and PathColumn syntax elements.

The function ignores input rows with NULL values.

InputTable Schema

Column Data Type Description
partition_column INTEGER, SHORT, or LONG Sequence index. Rows with same index belong to same sequence.
item_column CHARACTER or VARCHAR [Column appears only if you omit ItemDefinitionTable.] Sequence item.
time_column Any except DOUBLE PRECISION [Optional] Sequence item timestamp. Items in same sequence with same time stamp belong to same item set.
path_column CHARACTER or VARCHAR [Optional] Paths (as sequence strings).

ItemDefinitionTable Schema

Column Data Type Description
index_column INTEGER, SHORT, or LONG Used to determine when more than one predicate in definition_column is satisfied, in which case item_column applies.
definition_column CHARACTER or VARCHAR Predicate definition.
item_column CHARACTER or VARCHAR Sequence item for which predicate is true.

Function applies predicates to InputTable in index order. If more than one predicate is true for a row, function assigns row the value that corresponds to predicate with smallest index.

If InputTable row has no corresponding definition in item definition table, function skips that row.