计算在Python中有多少id报告给一个管理器

Question

我有一个使用python和numpy的数据集，我有一个数据集，看员工和他们的经理。我想要的是计算一个经理有多少份报告。

下面是样本数据集:

ID     Date          Job          Dept.     Manager ID 
1    Oct 2022     Sales Rep      Sales         5
1    Dec 2022     Sales Rep      Sales         5
1    Feb 2023     Sales Rep      Sales         5
2    Feb 2022     Tech Support   Tech          4
2    Jun 2022     Sales Advisor  Sales         5
2    Nov 2022     Sales Advisor  Sales         5
3    Dec 2021     Tech Consult   Tech          4 
3    Sept 2022    Tech Advisor   Tech          4

我想要的输出是:

Manager ID    Reports
4            1
5            2

我目前使用的代码是:

counts = df['ID'].groupby(df['Manager ID'].astype(float)).nunique()
df['Reports'] = df['ID'].astype(str).map(counts).fillna(0, downcast ='infer')

这段代码只输出零。对此有什么建议吗?

Answer 1

您可以尝试:

>>> (df.drop_duplicates(['ID', 'Manager ID'])
.groupby('Manager ID', as_index=False)
.agg(Report=('ID', 'size')))
Manager ID  Report
0           4       2
1           5       2

Answer 2

你只需要:

df.groupby('Manager ID')['ID'].nunique()

输出:

Manager ID
4    2
5    2

Answer 3

另一个可能的解决方案:

managers = df['Manager ID'].unique()
pd.DataFrame({
'Manager ID': managers,
'Reports': [df.loc[df['Manager ID'].eq(x), 'ID'].nunique() for x in managers]
})

输出:

Manager ID  Reports
0           5        2
1           4        2

Answer 4

似乎你可以这样做

import collections
print(collections.Counter(df['Manager ID']))