ValueError：：无法从重复轴重新索引

Question

我正在尝试用python编写一个函数。但是，它不断返回错误“无法从重复轴重新索引”。我不知道为什么我会出错

def get_match_digital(Demo, Imp, imp_cap):
    df102 = []
    # Loop for running a simulation n times
    d1 = df_new
    impcap = imp_cap

    for x in range(1):
        # Randomly selecting n number of random rows which would be no. of impressions in this case

        f2 = d1.sample(Imp)
        # applying frequency capping after sample selection
        d = f2.assign(rn=f2.sort_values(['MemberId'], ascending=False).groupby(['MemberId']).cumcount() + 1).query('rn <= 3')

        a = 20000
        df42 = []
        for y in range(10):
            df = f2.iloc[:a]

            df32 = df.loc[(df['ageGroup25_54'] == Demo)].MemberId.nunique()

            a = a + a

            df42.append(df32)

        df102.append(df42)

    transposed2 = list(zip(*df102))
    avg2 = lambda items: float(sum(items)) / len(items)

    averages2 = list(map(avg2, transposed2))
    columns = ['reach']
    final = pd.DataFrame(columns=columns)
    final = final.assign(reach=averages2)
    final['Imp'] = 20000 * (final.index.values + 1)
    final['reachP'] = round(((final['reach'] / Imp) * 100), 2)
    return final

我认为给出错误的代码行是

d = f2.assign(rn=f2.sort_values(['MemberId'], ascending=False).groupby(['MemberId']).cumcount() + 1).query('rn <= 3')

我在这里要做的是确保在随机样本选择中相同memberid的出现次数不超过3次

一旦我们删除这行代码，函数就可以工作了。我们试过用

d = f2.assign(rn=f2.sort_values(['MemberId'], ascending=False).groupby(level=0)(['MemberId']).cumcount() + 1).query('rn <= 3')

通过重置索引

d.reset_index()

我希望有人能帮我找出错误的原因，帮助我解决问题

提前多谢