TypeError: len() of unsized object

• 在使用pandas时，使用groupby对DataFrame对象分组，出现了该错误
• 堆栈信息：

Traceback (most recent call last):
  File "F:/workspace/sign_rate 4.0/machine_learning/sign_rate.py", line 364, in <module>
    p.predict(begin_date="2019-10-01")
  File "F:/workspace/sign_rate 4.0/machine_learning/sign_rate.py", line 196, in predict
    return self.get_complete_proportion_sign_rate(subdivide, mode)
  File "F:/workspace/sign_rate 4.0/machine_learning/sign_rate.py", line 228, in get_complete_proportion_sign_rate
    return self.get_month_cpsr(self.df, subdivide, mode)
  File "F:/workspace/sign_rate 4.0/machine_learning/sign_rate.py", line 248, in get_month_cpsr
    return self.get_category_cpsr(df, subdivide, mode)
  File "F:/workspace/sign_rate 4.0/machine_learning/sign_rate.py", line 285, in get_category_cpsr
    for cate, df_cate in df_area.groupby(by=subdivide):
  File "F:\envs\data_analysis\lib\site-packages\pandas\core\generic.py", line 7894, in groupby
    **kwargs
  File "F:\envs\data_analysis\lib\site-packages\pandas\core\groupby\groupby.py", line 2522, in groupby
    return klass(obj, by, **kwds)
  File "F:\envs\data_analysis\lib\site-packages\pandas\core\groupby\groupby.py", line 391, in __init__
    mutated=self.mutated,
  File "F:\envs\data_analysis\lib\site-packages\pandas\core\groupby\grouper.py", line 652, in _get_grouper
    if not isinstance(gpr, Grouping)
  File "F:\envs\data_analysis\lib\site-packages\pandas\core\groupby\grouper.py", line 259, in __init__
    self.grouper = _convert_grouper(index, grouper)
  File "F:\envs\data_analysis\lib\site-packages\pandas\core\groupby\grouper.py", line 681, in _convert_grouper
    if len(grouper) != len(axis):
TypeError: len() of unsized object

• 主要错误代码：

subdivide = np.where(subdivide is None, DefaultArgs.ID, subdivide)
for area, df_area in df.groupby(DefaultArgs.AREA):
     for cate, df_cate in df_area.groupby(subdivide):
     ...

• 原因：
  groupby可接受的参数类型有：mapping, function, label, or list of labels，一般只用单个字符串或者多个字符串列表用于分组。代码中DefaultArgs.ID subdivide均为字符串，按说都不应该出现问题，然而却出现了该问题。究其原因：np.where返回的对象不是想象中的str格式，返回的是 out : ndarray，本以为np.where的效率较if...else的效率会高不少，却出现预计之外的错误。
• 修改：

subdivide = DefaultArgs.ID if subdivide is None else subdivide