Spark3：pyspark注册udf和使用窗口函数

一、准备数据

%pyspark
data = [(1,"火男",6000),(1,"流浪法师",7000),(2,"盖伦",10000),(2,"皇子",8000)]
df = spark.createDataFrame(data,["属性","角色名称","战斗力"])
df.show()

二、pyspark注册udf

%pyspark
#注册udf
import pandas as pd
from pyspark.sql.functions import pandas_udf
from pyspark.sql.types import IntegerType

@pandas_udf("double")
def mean_udf(v: pd.Series) -> float:
    return v.mean()

<