PySpark:将参数转换为月份的最后一个工作日

from pyspark.sql.functions import col, last_day, expr, when
from pyspark.sql import SparkSession

spark = SparkSession.builder.appName("PySparkExample").getOrCreate()

data = [("2022-01-01"), ("2022-02-15"), ("2022-03-31")]
df = spark.createDataFrame(data, ["date"])

df = df.withColumn("date", col("date").cast("date"))

df = df.withColumn("last_day", last_day(col("date")))

df = df.withColumn("last_workday", when(expr("day(last_day)") >= 5, 
                                          expr("date_sub(last_day, 
                                          expr('case when dayofweek(last_day) >= 5 then dayofweek(last_day) - 5 else dayofweek(last_day) + 2 end)'))")
                                      .otherwise(expr("date_sub(last_day, 
                                          expr('case when dayofweek(last_day) = 1 then 2 else 1 end)')))))

df.show()