Question

我已经在这里检查了所有建议的答案，但是我尝试的每个答案我都认为我会破坏代码。

我的数据示例：

print(transactions.head())

                      loc  amount  local_date
0                    RAIL     8.1  2016-09-30
1      LINK NETWORK LIMIT     4.0  2016-10-02
2          CHOCOLATE CAFE     3.0  2016-10-03
3         Four Star Pizza     9.7  2016-10-03
4                  Cinema    10.0  2016-10-04

我只想按年份，月份和交易总额进行分组。

例如我的预期结果：

2019 Jan  100
     Feb  123 
     Mar  150
     etc.
2018 Jan  200
     Feb  150
     Mar  211
     etc.

我尝试过的（基本上是所有建议的答案）

transactions.set_index('local_date').groupby([(transactions.index.year),(transactions.index.month)])['amount'].sum()


AttributeError                            Traceback (most recent call last)
<ipython-input-332-64938cfdee85> in <module>
----> 1 transactions.set_index('local_date').groupby([(transactions.index.year),(transactions.index.month)])['amount'].sum()

AttributeError: 'RangeIndex' object has no attribute 'year'

transactions.set_index('local_date').groupby([(transactions.index.dt.year),(transactions.index.dt.month)])['amount'].sum()

AttributeError                            Traceback (most recent call last)
<ipython-input-334-150c05241676> in <module>
----> 1 transactions.set_index('local_date').groupby([(transactions.index.dt.year),(transactions.index.dt.month)])['amount'].sum()


AttributeError: 'RangeIndex' object has no attribute 'dt'

transactions.set_index('local_date').groupby([(transactions.index.to_series().dt.year),(transactions.index.to_series.()dt.month)])['amount'].sum()


AttributeError: Can only use .dt accessor with datetimelike values

我迷失了方向。我要去哪里错了？

Answer 1

使用Series.dt + DataFrame.groupby：

df['local_date']=pd.to_datetime(df['local_date'])
df.groupby([df['local_date'].dt.year,df['local_date'].dt.month])['amount'].sum()

local_date  local_date
2016        9              8.1
            10            26.7

如果愿意，请显示月份名称：

new_df=df.groupby([df['local_date'].dt.year,df['local_date'].dt.month_name()])['amount'].sum().to_frame('Total amount')
print(new_df)
                       Total amount
local_date local_date              
2016       October             26.7
           September            8.1

new_df=df.groupby(df['local_date'].dt.to_period('M')).amount.sum().to_frame('Total_amount')
print(new_df)

            Total_amount
local_date              
2016-09              8.1
2016-10             26.7

Answer 2

您可以按year, month分组：

(transaction.groupby([transaction.local_date.dt.year,
                     transaction.local_date.dt.month])
          .sum())

输出：

                       amount
local_date local_date        
2016       9              8.1
           10            26.7

如果要使用月份名称，请将.dt.month替换为.dt.month_name()，但是您需要做更多的工作才能正确排序：

                       amount
local_date local_date        
2016       October       26.7
           September      8.1

如何按年份和月份对交易进行分组？

2 个答案: