Python Pandas 中的 Parse_dates
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/23797491/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Parse_dates in Pandas
提问by user3576212
The following code can't parse my date column into dates from csv file.
以下代码无法将我的日期列解析为 csv 文件中的日期。
data=pd.read_csv('c:/data.csv',parse_dates=True,keep_date_col = True)
or
或者
data=pd.read_csv('c:/data.csv',parse_dates=[0])
data is like following
数据如下
date value
30MAR1990 140000
30JUN1990 30000
30SEP1990 120000
30DEC1990 34555
What did I do wrong? Please help!
我做错了什么？请帮忙！
Thanks.
谢谢。
采纳答案by Andy Hayden
This is a non-standard format, so not caught by the default parser, you can pass your own:
这是一种非标准格式，因此不会被默认解析器捕获，您可以传递自己的：
In [11]: import datetime as dt
In [12]: dt.datetime.strptime('30MAR1990', '%d%b%Y')
Out[12]: datetime.datetime(1990, 3, 30, 0, 0)
In [13]: parser = lambda date: pd.datetime.strptime(date, '%d%b%Y')
In [14]: pd.read_csv(StringIO(s), parse_dates=[0], date_parser=parser)
Out[14]:
date value
0 1990-03-30 140000
1 1990-06-30 30000
2 1990-09-30 120000
3 1990-12-30 34555
Another option is to use to_datetime afteryou've read in the strings:
另一种选择是在读入字符串后使用 to_datetime ：
df['date'] = pd.to_datetime(df['date'], format='%d%b%Y')
回答by TomAugspurger
You can use the date_parserargument to read_csv
您可以使用date_parserread_csv的参数
In [62]: from pandas.compat import StringIO
In [63]: s = """date,value
30MAR1990,140000
30JUN1990,30000
30SEP1990,120000
30DEC1990,34555
"""
In [64]: from pandas.compat import StringIO
In [65]: import datetime
date_parserexpects a function that will be called on an array of strings. funccalls datetime.datetime.strptimeon each string. Check out the datetimemodule in the python docs for more on the format codes.
date_parser期望一个将在字符串数组上调用的函数。func调用datetime.datetime.strptime每个字符串。datetime有关格式代码的更多信息，请查看python 文档中的模块。
In [66]: func = lambda dates: [datetime.datetime.strptime(x, '%d%b%Y') for x in dates]
In [67]: s = """date,value
30MAR1990,140000
30JUN1990,30000
30SEP1990,120000
30DEC1990,34555
"""
In [68]: pd.read_csv(StringIO(s), parse_dates=['date'], date_parser=func)
Out[68]:
date value
0 1990-03-30 140000
1 1990-06-30 30000
2 1990-09-30 120000
3 1990-12-30 34555
[4 rows x 2 columns]
