pyspark.sql.functions的unix_timestamp仅对于大多数值都为null

Question

我尝试使用此代码将列从字符串转换为时间戳记

from pyspark.sql import SparkSession
from pyspark.sql.functions import sum, avg, count, first, col
from pyspark.sql.functions import unix_timestamp, col, datediff
import csv

if __name__ == "__main__":
    spark = SparkSession \
        .builder \
        .config("spark.some.config.option", "some-value") \
        .getOrCreate()
    df=spark.read.csv("in\SETP11.csv")
    df = df.selectExpr('_c0 as USER_ID', '_c1 as timestamp', '_c2 as location')
    df.show()
    df = df.withColumn('timestamp', unix_timestamp('timestamp', 'yyyy-mm-dd hh:mm:ss'))
    df.show()

但是我得到空值（该函数只为某些返回值）

+-------+---------+--------+
|USER_ID|timestamp|location|
+-------+---------+--------+
|      1|     null|     321|
|      1|1359262.2|     386|
|      1|1359271.2|     386|
|      1|1359273.6|     386|
|      1|1359273.6|     386|
|      1|1359274.8|     386|
|      1|1359281.4|     386|
|      1|1359282.0|     386|
|      1|     null|     386|
|      1|     null|     386|
|      1|     null|     386|
|      1|     null|     386|
+-------+---------+--------+

该如何解决？问题在哪里？

关键字：python apache-spark pyspark unix-timestamp