从包含Pyspark中时间戳的字符串列中提取日期 [英] Extract date from a string column containing timestamp in Pyspark

问题描述

我有一个数据框，其日期格式如下:

I have a dataframe which has a date in the following format:

+----------------------+
|date                  |
+----------------------+
|May 6, 2016 5:59:34 AM|
+----------------------+

我打算以YYYY-MM-DD格式从中提取日期；因此结果应为上述日期-2016-05-06.

I intend to extract the date from this in the format YYYY-MM-DD ; so the result should be for the above date - 2016-05-06.

但是当我提取时使用的是以下内容:

But when I extract is using the following:

df.withColumn('part_date', from_unixtime(unix_timestamp(df.date, "MMM dd, YYYY hh:mm:ss aa"), "yyyy-MM-dd"))

我得到以下日期

2015-12-27

有人可以对此提出建议吗?我不打算将df转换为rdd以使用来自python的datetime函数，并希望在自身的数据框中使用它.

Can anyone please advise on this? I do not intend to convert my df to rdd to use datetime function from python and want to use this in the dataframe it self.

推荐答案

您的模式存在一些错误.这是一个建议:

There are some errors with your pattern. Here's a suggestion:

from_pattern = 'MMM d, yyyy h:mm:ss aa'
to_pattern = 'yyyy-MM-dd'
df.withColumn('part_date', from_unixtime(unix_timestamp(df['date'], from_pattern), to_pattern)).show()

+----------------------+----------+
|date                  |part_date |
+----------------------+----------+
|May 6, 2016 5:59:34 AM|2016-05-06|
+----------------------+----------+

这篇关于从包含Pyspark中时间戳的字符串列中提取日期的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！