Question

我有一个数据集，其中一列称为“ YearMade”，其类型为int64。我正在尝试替换“ YearMade”列中的值，其中任何小于1918的值都将替换为该列的中位数。所以我尝试：

df.where(df['YearMade'] > 1918, df['YearMade'].median(), inplace = True)

但是，我遇到类型错误。我在这里做什么错？而且，如何纠正呢？请参阅下面的错误消息：

<ipython-input-83-b202aa389b1d> in <module>
      1 # We replace all the rows before 1929 with the median
      2 
----> 3 df.where(df['YearMade'] > 1918, df['YearMade'].median(), inplace = True)
      4 df['YearMade'].describe()

/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py in where(self, cond, other, inplace, axis, level, errors, try_cast)
   9274         other = com.apply_if_callable(other, self)
   9275         return self._where(
-> 9276             cond, other, inplace, axis, level, errors=errors, try_cast=try_cast
   9277         )
   9278 

/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py in _where(self, cond, other, inplace, axis, level, errors, try_cast)
   9103             # reconstruct the block manager
   9104 
-> 9105             self._check_inplace_setting(other)
   9106             new_data = self._data.putmask(
   9107                 mask=cond,

/opt/anaconda3/lib/python3.7/site-packages/pandas/core/generic.py in _check_inplace_setting(self, value)
   5303 
   5304                 raise TypeError(
-> 5305                     "Cannot do inplace boolean setting on "
   5306                     "mixed-types with a non np.nan value"
   5307                 )

TypeError: Cannot do inplace boolean setting on mixed-types with a non np.nan value```

Answer 1

IIUC，您只想替换特定列的值。我认为您会收到错误，因为where条件应用于所有具有dtypes混合的列。试试这个：

df['YearMade'].where(df['YearMade'] > 1918, df['YearMade'].median(), inplace = True)

使用熊猫“ where”功能时出现类型错误

1 个答案: