Pandas升级至2.1.1后pivot_table报TypeError问题求助

问题背景 #

将Pandas从1.3.2版本升级到2.1.1版本后，以下代码运行时抛出TypeError: super(type, obj): obj must be an instance or subtype of type错误：

报错代码片段 #

import pandas as pd

d = {'value': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0],
     'x_axis_date': ['2020-01-01', '2020-01-02', '2020-01-03', '2020-01-04', '2020-01-05', '2020-01-06', '2020-01-07',
                     '2020-01-08', '2020-01-09', '2020-01-10'],
     'curve_year': [2019, 2019, 2019, 2019, 2019, 2019, 2019, 2019, 2019, 2019]}
df = pd.DataFrame(data=d)
df = df.assign(x_axis_date=pd.to_datetime(df['x_axis_date']))

# dtypes: float64, datetime64[ns], int64
pv_aggregate_agg = pd.pivot_table(df, index=df.x_axis_date, values=['value'], aggfunc=pd.DataFrame.mean) # error here
print(pv_aggregate_agg)

堆栈跟踪 #

Traceback (most recent call last):
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\generic.py", line 1495, in aggregate
    result = gba.agg()
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\apply.py", line 178, in agg
    return self.agg_list_like()
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\apply.py", line 311, in agg_list_like
    return self.agg_or_apply_list_like(op_name="agg")
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\apply.py", line 1351, in agg_or_apply_list_like
    keys, results = self.compute_list_like(op_name, selected_obj, kwargs)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\apply.py", line 370, in compute_list_like
    new_res = getattr(colg, op_name)(func, *args, **kwargs)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\generic.py", line 255, in aggregate
    ret = self._aggregate_multiple_funcs(func, *args, **kwargs)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\generic.py", line 360, in _aggregate_multiple_funcs
    results[key] = self.aggregate(func, *args, **kwargs)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\generic.py", line 292, in aggregate
    return self._python_agg_general(func, *args, **kwargs)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\generic.py", line 325, in _python_agg_general
    result = self.grouper.agg_series(obj, f)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\ops.py", line 850, in agg_series
    result = self._aggregate_series_pure_python(obj, func)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\ops.py", line 871, in _aggregate_series_pure_python
    res = func(group)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\groupby\generic.py", line 322, in <lambda>
    f = lambda x: func(x, *args, **kwargs)
  File "E:\repos\plot_graphs\venv\lib\site-packages\pandas\core\frame.py", line 11335, in mean
    result = super().mean(axis, skipna, numeric_only, **kwargs)
TypeError: super(type, obj): obj must be an instance or subtype of type

原因分析 #

报错根源在于pd.pivot_table的aggfunc参数使用了pd.DataFrame.mean：

• 当values参数指定单列（如这里的['value']）时，pivot_table会对每个分组的Series对象应用聚合函数
• pd.DataFrame.mean是DataFrame类的方法，内部通过super()调用父类方法时，要求调用对象必须是DataFrame实例或其子类
• Pandas 1.3.2对这种类型不匹配的情况兼容性较强，而2.1.1版本严格了类型检查，因此触发错误

解决方法 #

方法1：使用Series的mean方法（推荐） #

将aggfunc改为pd.Series.mean，或者更简洁的字符串'mean'：

# 写法1：使用Series.mean
pv_aggregate_agg = pd.pivot_table(df, index=df.x_axis_date, values=['value'], aggfunc=pd.Series.mean)

# 写法2：直接用字符串'mean'（更简洁，pandas会自动匹配对应方法）
pv_aggregate_agg = pd.pivot_table(df, index=df.x_axis_date, values=['value'], aggfunc='mean')

方法2：自定义兼容DataFrame的聚合函数（适用于多列场景） #

如果需要处理多列聚合，确保函数能接收DataFrame对象：

def custom_mean(df):
    return df.mean()

pv_aggregate_agg = pd.pivot_table(df, index=df.x_axis_date, values=['value'], aggfunc=custom_mean)

修改后运行结果 #

两种方法都能正常输出预期结果：

value
x_axis_date       
2020-01-01    1.0
2020-01-02    1.0
2020-01-03    1.0
2020-01-04    1.0
2020-01-05    1.0
2020-01-06    1.0
2020-01-07    1.0
2020-01-08    1.0
2020-01-09    1.0
2020-01-10    1.0

内容的提问来源于stack exchange，提问作者Vitamin C