TypeError：无法连接“str”和“float”对象：pandas

Question

我正在尝试使用 pandas 解决数据科学问题。我的数据集包含以下列：“国家”、“转换”、“测试”、“用户 ID”等。在国家列中，大约有 10 个国家。 “测试”列的值 0 和 1 表示两种类型的测试：控制 0 和实验 1。“转换”也有值 0 和 1，表示该人是否已转换。

我想按国家/地区分组并计算每个组的 p 值和 test==0 和 test==1 的平均值。我正在尝试使用以下函数，但是它会引发错误，“TypeError: cannot concatenate 'str' and 'float' objects”。有人可以解释一下吗？

def f(x):
        control = x.loc[(x.test==0)]
        test = x.loc[(x.test==1)]
        p_value = stats.ttest_ind(control,test)[0]
        control_mean = control['conversion'].mean()
        test_mean = test['conversion'].mean()
        return pd.Series({'p_value': p_value, 'conversion_test': test_mean, 'conversion_control': control_mean})     

bycountry = data1.groupby('country').apply(f) 
bycountry = bycountry.reset_index(level='None')
bycountry

完整的错误信息：

TypeError                                 Traceback (most recent call last)
<ipython-input-495-bd6227878520> in <module>()
      7     return pd.Series({'p_value': p_value, 'conversion_test': test_mean, 'conversion_control': control_mean})
      8 
----> 9 bycountry = data1.groupby("country").apply(f)
     10 bycountry = bycountry.reset_index(level='None')
     11 bycountry

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\pandas\core\groupby.pyc in apply(self, func, *args, **kwargs)
    649         # ignore SettingWithCopy here in case the user mutates
    650         with option_context('mode.chained_assignment', None):
--> 651             return self._python_apply_general(f)
    652 
    653     def _python_apply_general(self, f):

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\pandas\core\groupby.pyc in _python_apply_general(self, f)
    653     def _python_apply_general(self, f):
    654         keys, values, mutated = self.grouper.apply(f, self._selected_obj,
--> 655                                                    self.axis)
    656 
    657         return self._wrap_applied_output(

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\pandas\core\groupby.pyc in apply(self, f, data, axis)
   1525             # group might be modified
   1526             group_axes = _get_axes(group)
-> 1527             res = f(group)
   1528             if not _is_indexed_like(res, group_axes):
   1529                 mutated = True

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\pandas\core\groupby.pyc in f(g)
    645         @wraps(func)
    646         def f(g):
--> 647             return func(g, *args, **kwargs)
    648 
    649         # ignore SettingWithCopy here in case the user mutates

<ipython-input-495-bd6227878520> in f(x)
      2     control = x.loc[(x.test==0)]
      3     test = x.loc[(x.test==1)]
----> 4     p_value = stats.ttest_ind(control,test)[0]
      5     control_mean = control['conversion'].mean()
      6     test_mean = test['conversion'].mean()

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\scipy\stats\stats.pyc in ttest_ind(a, b, axis, equal_var, nan_policy)
   3865         return Ttest_indResult(np.nan, np.nan)
   3866 
-> 3867     v1 = np.var(a, axis, ddof=1)
   3868     v2 = np.var(b, axis, ddof=1)
   3869     n1 = a.shape[axis]

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\numpy\core\fromnumeric.pyc in var(a, axis, dtype, out, ddof, keepdims)
   3098 
   3099     return _methods._var(a, axis=axis, dtype=dtype, out=out, ddof=ddof,
-> 3100                          keepdims=keepdims)

C:\Users\SnehaPriya\Anaconda2\lib\site-packages\numpy\core\_methods.pyc in _var(a, axis, dtype, out, ddof, keepdims)
     89     # Note that if dtype is not of inexact type then arraymean will
     90     # not be either.
---> 91     arrmean = umr_sum(arr, axis, dtype, keepdims=True)
     92     if isinstance(arrmean, mu.ndarray):
     93         arrmean = um.true_divide(

TypeError: cannot concatenate 'str' and 'float' objects

df.dtypes 的输出：

user_id                      int64
date                datetime64[ns]
source                      object
device                      object
browser_language            object
ads_channel                 object
browser                     object
conversion                   int64
test                         int64
sex                         object
age                        float64
country                     object
dtype: object

Answer 1

def f(x):
    control = x.loc[(x.test==0)]
    control = control['conversion']
    test = x.loc[(x.test==1)]
    test = test['conversion']
    p_value = stats.ttest_ind(control,test)[0]
    control_mean = control.mean()
    test_mean = test.mean()
    return pd.Series({'p_value': p_value, 'conversion_test': test_mean, 'conversion_control': control_mean})     

这成功了！！再次感谢@juanpa.arrivillaga！！