User Defined Functions

The udf() decorator creates user-defined scalar functions for use with the DataFrame API. It supports plain functions, async functions, ScalarFunction subclasses, and callable classes. The udtf() decorator creates user-defined table functions that emit zero or more rows per input, for use with flat_map() and join_lateral().

Example:

>>> from pyflink.dataframe import udf, udtf, DataType
>>> @udf(return_dtype=DataType.string())
... def to_upper(s: str) -> str:
...     return s.upper()
...
>>> @udtf(return_dtype=DataType.string())
... def split_words(text: str):
...     for word in text.split():
...         yield word

Functions

udf()	Create a user-defined function for DataFrame operations.
udtf()	Create a user-defined table function for DataFrame operations.