python - pandas 使用多索引进行条件更新

Question

假设我有以下 DataFrame，在索引和列上都有多个索引。

 first   x    y
 second  m    n
A  B
A0 B0    0    0
   B1    0    0
A1 B0    0    0
   B1    0    0

我正在尝试根据条件更新值。条件将类似于:`规则:

[{condition:{'A':'A0,'B':'B0'},value:5},
  {condition:{'B':'B1'},value:3},
 .....]

我正在尝试寻找具有类似功能的东西

• 使用pandas.DataFrame.xs设置值:

  for each rule in rules:
      df.xs((conditions.values), level=[conditions.keys]) = value
  

• 将多个级别传递给 pandas.Index.get_level_values 以设置值:

  for each rule in rules:
      df.loc[df.index.get_level_values(conditions.keys) == [conditions.values] = value
  

结果应该是

 first   x    y
 second  m    n
A  B
A0 B0    5    5
   B1    3    3
A1 B0    0    0
   B1    3    3`

Answer 1

不幸的是，pandas 的 MultiIndex 中的字典的选择尚未 not supported ，所以需要定制function适合您:

rules = [{'condition':{'A':'A0','B':'B0'},'value':5},
         {'condition':{'B':'B1'},'value':3}]

for rule in rules:
    d = rule['condition']
    indexer = [d[name] if name in d else slice(None) for name in df.index.names]
    df.loc[tuple(indexer),] = rule['value']

print (df)
first   x  y
second  m  n
A  B        
A0 B0   5  5
   B1   3  3
A1 B0   0  0
   B1   3  3