gravitino.api.function.function.Function

class gravitino.api.function.function.Function

Bases: Auditable

An interface representing a user-defined function under a schema Namespace.

A function is a reusable computational unit that can be invoked within queries across different compute engines. Users can register a function in Gravitino to manage the function metadata and enable cross-engine function sharing. The typical use case is to define custom business logic once and reuse it across multiple compute engines like Spark, Trino, and AI engines.

A function is characterized by its name, type (scalar for row-by-row operations, aggregate for group operations, or table-valued for set-returning operations), whether it is deterministic, and its definitions that contain parameters, return type or columns (for table function), and implementations for different runtime engines.

__init__()

Methods

__init__()
audit_info()
comment()	Returns the optional comment of the function.
definitions()	Returns the definitions of the function.
deterministic()	Returns whether the function is deterministic.
function_type()	Returns the function type.
name()	Returns the function name.
supports_tags()	Return tag operations if the function supports tags.

comment() → str | None

Returns the optional comment of the function.

abstract definitions() → List[FunctionDefinition]

Returns the definitions of the function.

abstract deterministic() → bool

Returns whether the function is deterministic.

abstract function_type() → FunctionType

Returns the function type.

abstract name() → str

Returns the function name.

supports_tags() → SupportsTags

Return tag operations if the function supports tags.

Raises:

UnsupportedOperationException: If the function does not support tag operations.

Returns:

SupportsTags: The tag operations supported by the function.