在通过代码连接的两个表之间使用更新时出错

是以下 python pandas DataFrame：

If code列不是NaN和money列是NaN，我们更新值money and money_add从下表中。使用code and cod_t列作为键。

结果表的示例：

用户@jezrael，给了我以下问题的解决方案：

df1 = df1.drop_duplicates('cod_t').set_index('cod_t')
df = df.set_index(df['code'])
df.update(df1, overwrite=False)
df = df.reset_index(drop=True).reindex(df.columns, axis=1)

但这段代码给了我一个我不知道如何解决的错误：

TypeError: The DType <class 'numpy.dtype[timedelta64]'> could not be promoted by <class
'numpy.dtype[float64]'>. This means that no common DType exists for the given inputs. 
For example they cannot be stored in a single array unless the dtype is `object`. 
The full list of DTypes is: (<class 'numpy.dtype[timedelta64]'>, <class 'numpy.dtype[float64]'>)

// First DataFrame dtypes
ID                                 object
country                            object
code                               object
money                             float64
money_add                         float64
other                              object
time                      timedelta64[ns]
dtype: object
// Second DataFrame dtypes
cod_t                     object
money                      int64
money_add                  int64
dtype: object

如果您能帮助我解决错误，或者建议使用替代方法，我将不胜感激update.

Because DataFrame.update这里效果不好是替代方案 - 首先对第二个 DataFrame 中的新列使用左连接DataFrame.merge:

df2 = df.merge(df1.drop_duplicates('cod_t').rename(columns={'cod_t':'code'}), 
               on='code', 
               how='left',
               suffixes=('','_'))

print (df2)
       ID country    money  code  money_add         other            time  \
0  832932   Other      NaN   0.0        NaN  [N2, N2, N4] 0 days 01:37:00   
1  217#8#     NaN      NaN   NaN        NaN  [N1, N2, N3] 2 days 01:01:00   
2  1329T2  France  12131.0  20.0     3452.0      [N1, N1] 1 days 03:55:00   
3  124932  France      NaN  16.0        NaN          [N2] 0 days 01:28:00   
4  194022  France      NaN   0.0        NaN      [N4, N3] 3 days 02:35:00   

   money_  money_add_  
0  4532.0     72323.0  
1     NaN         NaN  
2  1813.0     27328.0  
3  1213.0     23822.0  
4  4532.0     72323.0 

然后获取带/不带的列名称_:

cols_with_ = df2.columns[df2.columns.str.endswith('_')]
cols_without_ = cols_with_.str.rstrip('_')

print (cols_with_)
Index(['money_', 'money_add_'], dtype='object')

print (cols_without_)
Index(['money', 'money_add'], dtype='object')

Pass to DataFrame.combine_first最后删除辅助列：

df2[cols_without_] = (df2[cols_without_].combine_first(df2[cols_with_]
                                        .rename(columns=lambda x: x.rstrip('_'))))
df2 = df2.drop(cols_with_, axis=1)
print (df2)
       ID country    money  code  money_add         other            time
0  832932   Other   4532.0   0.0    72323.0  [N2, N2, N4] 0 days 01:37:00
1  217#8#     NaN      NaN   NaN        NaN  [N1, N2, N3] 2 days 01:01:00
2  1329T2  France  12131.0  20.0     3452.0      [N1, N1] 1 days 03:55:00
3  124932  France   1213.0  16.0    23822.0          [N2] 0 days 01:28:00
4  194022  France   4532.0   0.0    72323.0      [N4, N3] 3 days 02:35:00