如何将Pandas Index转换为月份名称

Question

我试图将以下时间序列中的日期索引更改为月份名称。

website = dfFinal.groupby(['Date','Website'])

websiteGroup = website['Visits'].aggregate(np.sum).unstack()

Website              A        B          C
Date                                      
2015-01-01       18185   805769        NaN
2015-02-01       73236   944458        NaN
2015-03-01      101737  1003966        NaN
2015-04-01      101018   861229        NaN
2015-05-01       77724   845223        NaN
2015-06-01      111503   966043        NaN
2015-07-01      115413   937184        NaN
2015-08-01      115215   890457       1649

例如，我希望它看起来像这样：

    Website           A        B          C
    Date                                      
    January       18185   805769        NaN
    February      73236   944458        NaN
    March        101737  1003966        NaN
    April        101018   861229        NaN
    May           77724   845223        NaN
    June         111503   966043        NaN
    July         115413   937184        NaN
    August       115215   890457       1649

我希望能够这样，所以我的情节标记将是月份名称而不是日期时间。

谢谢

编辑//

相同的方案，但解决方案不起作用：

systemType = dfFinal.groupby(['Date','Website','Type'])
systemGroup = systemType['Visits'].aggregate(np.sum)
systemGroup = systemGroup.groupby(level=[0,1]).apply(lambda x: 100*x/float(x.sum())).unstack()

Type                      Other  Windows Mobile  Windows PC  
Date       Website                                           
2015-01-01 A           0.637888        0.005499   48.814957  
           B           0.686549        0.016506   54.176073  
2015-02-01 A           0.742804        0.020482   49.811568  
           B           0.651802        0.014506   57.014288  
2015-03-01 A           0.668390        0.014744   50.087972  
           B           0.573924        0.015937   59.906013  
2015-04-01 A           0.662258        0.015839   49.310024  
           B           0.583933        0.013469   59.490449  
2015-05-01 A           0.666461        0.020586   48.522979  
           B           0.577954        0.017983   58.838200  


systemGroup = systemGroup.rename(index=lambda x: x.strftime('%B'))

给了我一个错误

AttributeError: 'str' object has no attribute 'strftime'

Answer 1

如果您有DatetimeIndex，则可以使用

websiteGroup.rename(index=lambda x: x.strftime('%B'))

.rename可以使用函数，我们将使用'%B'代码作为完整的月份名称。

Answer 2

使用DatetimeIndex.strftime：

websiteGroup.index = websiteGroup.index.strftime('%B')
print (websiteGroup)
               A        B       C
January    18185   805769     NaN
February   73236   944458     NaN
March     101737  1003966     NaN
April     101018   861229     NaN
May        77724   845223     NaN
June      111503   966043     NaN
July      115413   937184     NaN
August    115215   890457  1649.0

---

df = websiteGroup.set_index(websiteGroup.index.strftime('%b'))
print (df)
          A        B       C
Jan   18185   805769     NaN
Feb   73236   944458     NaN
Mar  101737  1003966     NaN
Apr  101018   861229     NaN
May   77724   845223     NaN
Jun  111503   966043     NaN
Jul  115413   937184     NaN
Aug  115215   890457  1649.0

另外，为了在索引中分配新值，可以使用set_index：

df = websiteGroup.set_index(websiteGroup.index.strftime('%B'))
print (df)
               A        B       C
January    18185   805769     NaN
February   73236   944458     NaN
March     101737  1003966     NaN
April     101018   861229     NaN
May        77724   845223     NaN
June      111503   966043     NaN
July      115413   937184     NaN
August    115215   890457  1649.0

编辑：

对于版本pandas 0.23.0可能使用DatetimeIndex.month_name：

websiteGroup.index = websiteGroup.index.month_name()
print (websiteGroup)
               A        B       C
Website                          
January    18185   805769     NaN
February   73236   944458     NaN
March     101737  1003966     NaN
April     101018   861229     NaN
May        77724   845223     NaN
June      111503   966043     NaN
July      115413   937184     NaN
August    115215   890457  1649.0

Answer 3

您可以使用datetime.strptime解析每个日期字符串，并使用datetime.strftime('%B')打印月份名称：

>>> d = datetime.datetime.strptime('2015-01-01', '%Y-%m-%d')
>>> d.strftime('%B')
'January'

Answer 4

Pandas版本0.23.0及以后（截至本文时，它为0.24.2）提供了一种内置方法：.month_name。来自its official documentation：

pandas.DatetimeIndex.month_name返回具有指定语言环境的DateTimeIndex的月份名称。

考虑以下DataFrame：

aapl.tail()
# returns:
Attributes  High    Low Open    Close   Volume  Adj Close
2019-03-27  189.76  186.55  188.75  188.47  29848400.0  188.47
2019-03-28  189.56  187.53  188.95  188.72  20780400.0  188.72
2019-03-29  190.08  188.54  189.83  189.95  23564000.0  189.95
2019-03-30  190.08  188.54  189.83  189.95  23564000.0  189.95
2019-03-31  190.08  188.54  189.83  189.95  23564000.0  189.95

DataFrame有一个DateTimeIndex，所以我们可以在索引上应用.month_name，如下所示：

aapl.index = aapl.index.month_name()
aapl.tail()
# returns:
Attributes  High    Low Open    Close   Volume  Adj Close
March   189.76  186.55  188.75  188.47  29848400.0  188.47
March   189.56  187.53  188.95  188.72  20780400.0  188.72
March   190.08  188.54  189.83  189.95  23564000.0  189.95
March   190.08  188.54  189.83  189.95  23564000.0  189.95
March   190.08  188.54  189.83  189.95  23564000.0  189.95

在0.23.0之前你会使用.month()和其他参考使用.strftime('%B')的答案是要走的路。