python-3.x - 无法使用带有标记为索引的字符串的 loc 进行设置

Question

我使用的是 pandas v23.4 和 python 3.7.0。我有两个具有相同形状和列的表。我想将一个表的索引选择子集设置为另一个表的相同索引选择子集。

它有时无法使用字符串索引，但我不确定它是否与它是一个字符串索引有关。注释掉数据框的一列未使用的列会使其正常工作。

下面的回溯显示它在索引中的某个地方，在确定目标和源是否具有相同长度时会感到困惑。代码很长，有点绕。

$ cat foo.py
import numpy as np
import pandas as pd

m = np.array([1., 2., 1., 3., 5., 5., 6., 2., 2., 1., 7., 2.,
              5., 4., 2., 5., 5., 5., 3., 8., 7., 2., 7., 6.], )
dma_l = [501, 501, 501, 501, 501, 501, 501, 501, 501, 501, 501, 501,
         502, 502, 502, 502, 502, 502, 502, 502, 502, 502, 502, 502]
size_l = [1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4,
          1, 1, 1, 2, 2, 2, 3, 3, 3, 4, 4, 4]

age_l = ['20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45',
         '20-25', '30-35', '40-45']

df = pd.DataFrame()
df['dma'] = dma_l  # <-- comment out this line and it works
df['size'] = size_l
df['age'] = age_l
df['total'] = m

df2 = df.copy() # Make a second dataframe with the same shape.

# Works with an integer index.
df.set_index('size', inplace=True)
df2.set_index('size', inplace=True)
df.loc[(1,), 'total'] = df2.loc[(1,), 'total']

# Does not work with my string index.  Removing the dma column 
# causes it to work again.    
df.set_index('age', inplace=True)
df2.set_index('age', inplace=True)
df.loc[('20-25',), 'total'] = df2.loc[('20-25',), 'total']

$ python foo.py
Traceback (most recent call last):
  File "...", line 34, in <module>
    df.loc[('20-25',), 'total'] = df2.loc[('20-25',), 'total']
  File ".../lib/python3.7/site-packages/pandas/core/indexing.py", line 189, in __setitem__
    self._setitem_with_indexer(indexer, value)
  File ".../lib/python3.7/site-packages/pandas/core/indexing.py", line 606, in _setitem_with_indexer
    raise ValueError('Must have equal len keys and value '
ValueError: Must have equal len keys and value when setting with an iterable

Answer 1

如果两个 DataFrame 中的行数和索引值相同，则所有解决方案都有效。

---

问题是重复的索引值，如果想使用一个索引值，解决方案是创建 bool 掩码:

df2['total'] *= 10
df.loc[df.index == 1, 'total'] = df2.loc[1, 'total']
print (df)
      dma    age  total
size                   
1     501  20-25   10.0
1     501  30-35   20.0
1     501  40-45   10.0
2     501  20-25    3.0
2     501  30-35    5.0
2     501  40-45    5.0
3     501  20-25    6.0
3     501  30-35    2.0
3     501  40-45    2.0
4     501  20-25    1.0
4     501  30-35    7.0
4     501  40-45    2.0
1     502  20-25   50.0
1     502  30-35   40.0
1     502  40-45   20.0
2     502  20-25    5.0
2     502  30-35    5.0
2     502  40-45    5.0
3     502  20-25    3.0
3     502  30-35    8.0
3     502  40-45    7.0
4     502  20-25    2.0
4     502  30-35    7.0
4     502  40-45    6.0

或者通过分配创建新列以获得更通用的解决方案:

df2['total'] *= 10
df['total1'] = df2['total']
#working with one DataFrame
df.loc[[1, 4], 'total'] = df.loc[[1, 4], 'total1']
print (df)
      dma    age  total  total1
size                           
1     501  20-25   10.0    10.0
1     501  30-35   20.0    20.0
1     501  40-45   10.0    10.0
2     501  20-25    3.0    30.0
2     501  30-35    5.0    50.0
2     501  40-45    5.0    50.0
3     501  20-25    6.0    60.0
3     501  30-35    2.0    20.0
3     501  40-45    2.0    20.0
4     501  20-25   10.0    10.0
4     501  30-35   70.0    70.0
4     501  40-45   20.0    20.0
1     502  20-25   50.0    50.0
1     502  30-35   40.0    40.0
1     502  40-45   20.0    20.0
2     502  20-25    5.0    50.0
2     502  30-35    5.0    50.0
2     502  40-45    5.0    50.0
3     502  20-25    3.0    30.0
3     502  30-35    8.0    80.0
3     502  40-45    7.0    70.0
4     502  20-25   20.0    20.0
4     502  30-35   70.0    70.0
4     502  40-45   60.0    60.0

另一种解决方案是在两个 DataFrame 中创建掩码和过滤器:

df2['total'] *= 10

mask = df.index.isin([1,4])
df.loc[mask, 'total'] = df2.loc[mask, 'total']
print (df)
      dma    age  total
size                   
1     501  20-25   10.0
1     501  30-35   20.0
1     501  40-45   10.0
2     501  20-25    3.0
2     501  30-35    5.0
2     501  40-45    5.0
3     501  20-25    6.0
3     501  30-35    2.0
3     501  40-45    2.0
4     501  20-25   10.0
4     501  30-35   70.0
4     501  40-45   20.0
1     502  20-25   50.0
1     502  30-35   40.0
1     502  40-45   20.0
2     502  20-25    5.0
2     502  30-35    5.0
2     502  40-45    5.0
3     502  20-25    3.0
3     502  30-35    8.0
3     502  40-45    7.0
4     502  20-25   20.0
4     502  30-35   70.0
4     502  40-45   60.0