Del via

register (UDFRegistration)

Registers a Python function (including lambda functions) or a user-defined function as a SQL function.

Syntax

register(name, f, returnType=None)

Parameters

Parameter Type Description
name str Name of the user-defined function in SQL statements.
f function, udf, or pandas_udf A Python function, or a user-defined function. The user-defined function can be either row-at-a-time or vectorized.
returnType DataType or str, optional The return type of the registered user-defined function. Can be a DataType object or a DDL-formatted type string. Only valid when f is a plain Python function, not when f is already a user-defined function.

Returns

function

Notes

To register a nondeterministic Python function, first build a nondeterministic user-defined function for the Python function and then register it as a SQL function.

Examples

# Register a lambda as a SQL function (return type defaults to string).
strlen = spark.udf.register("stringLengthString", lambda x: len(x))
spark.sql("SELECT stringLengthString('test')").collect()
# [Row(stringLengthString(test)='4')]

spark.sql("SELECT 'foo' AS text").select(strlen("text")).collect()
# [Row(stringLengthString(text)='3')]

# Register with an explicit return type.
from pyspark.sql.types import IntegerType
spark.udf.register("stringLengthInt", lambda x: len(x), IntegerType())
spark.sql("SELECT stringLengthInt('test')").collect()
# [Row(stringLengthInt(test)=4)]

# Register an existing UDF.
from pyspark.sql.functions import udf
slen = udf(lambda s: len(s), IntegerType())
spark.udf.register("slen", slen)
spark.sql("SELECT slen('test')").collect()
# [Row(slen(test)=4)]

# Register a nondeterministic UDF.
import random
random_udf = udf(lambda: random.randint(0, 100), IntegerType()).asNondeterministic()
spark.udf.register("random_udf", random_udf)

# Register a pandas UDF.
import pandas as pd
from pyspark.sql.functions import pandas_udf

@pandas_udf("integer")
def add_one(s: pd.Series) -> pd.Series:
    return s + 1

spark.udf.register("add_one", add_one)
spark.sql("SELECT add_one(id) FROM range(3)").collect()
# [Row(add_one(id)=1), Row(add_one(id)=2), Row(add_one(id)=3)]

# Register a grouped aggregate pandas UDF.
@pandas_udf("integer")
def sum_udf(v: pd.Series) -> int:
    return v.sum()

spark.udf.register("sum_udf", sum_udf)
spark.sql(
    "SELECT sum_udf(v1) FROM VALUES (3, 0), (2, 0), (1, 1) tbl(v1, v2) GROUP BY v2"
).sort("sum_udf(v1)").collect()
# [Row(sum_udf(v1)=1), Row(sum_udf(v1)=5)]
  • Last updated on