在pyspark sql中查找两个时间戳之间的差异

from pyspark.sql import SparkSession
from pyspark.sql.functions import to_date, datediff

# 创建SparkSession
spark = SparkSession.builder.getOrCreate()

# 创建示例数据
data = [("2022-01-01 10:00:00", "2022-01-03 12:00:00"),
        ("2022-02-01 08:00:00", "2022-02-01 18:00:00"),
        ("2022-03-15 15:30:00", "2022-03-16 09:30:00")]

df = spark.createDataFrame(data, ["start_time", "end_time"])

# 将时间戳转换为日期类型
df = df.withColumn("start_date", to_date(df.start_time))
df = df.withColumn("end_date", to_date(df.end_time))

# 计算日期差异
df = df.withColumn("date_diff", datediff(df.end_date, df.start_date))

# 显示结果
df.show()

+-------------------+-------------------+----------+----------+---------+
|         start_time|           end_time|start_date|  end_date|date_diff|
+-------------------+-------------------+----------+----------+---------+
|2022-01-01 10:00:00|2022-01-03 12:00:00|2022-01-01|2022-01-03|        2|
|2022-02-01 08:00:00|2022-02-01 18:00:00|2022-02-01|2022-02-01|        0|
|2022-03-15 15:30:00|2022-03-16 09:30:00|2022-03-15|2022-03-16|        1|
+-------------------+-------------------+----------+----------+---------+