将字符串转换为日期列

Question

我想将字符串列转换为日期列，但结果我收到了包含 null 值的列。

from pyspark.sql.functions import expr, from_unixtime, dayofmonth, unix_timestamp, year, to_date, col

dane3.withColumn('date', to_date(unix_timestamp(col('dateRep'),'%d.%m.%Y').cast("timestamp"))).show()

Answer 1

您不需要像 sql 这样的d/m/y之前的% 。文档中的更多信息

df = spark.createDataFrame(
    [
        (1, '27.08.2020'), 
        (2, '27.08.2019'),
    ],
    ['id', 'txt']
)

df = df.withColumn('formatted', 
                   to_date(unix_timestamp(col('txt'), 'dd.MM.yyyy').cast("timestamp")))
df.show()

+---+----------+----------+
| id|       txt| formatted|
+---+----------+----------+
|  1|27.08.2020|2019-12-29|
|  2|27.08.2019|2018-12-30|
+---+----------+----------+