熊猫按列值分组并获取计数

Question

我有结果、学生、版本和状态列。在这个我想通过使用 Student ， Version 和 result = pass count 和 result = fail count 来分组

类似于 df.groupby(["student", "version", "result=pass"]).size().reset_index(name="new_result")

下面是我的数据框

result	student	version	status
pass	Student-A	L-1.0	Active
fail	Student-A	L-1.0	Active
fail	Student-A	L-1.0	Active
pass	Student-A	M-1.0	Active
fail	Student-A	M-1.0	Active
pass	Student-B	N-1.0	Active
pass	Student-B	N-1.0	Active
pass	Student-B	N-1.0	Active
pass	Student-C	O-1.0	Active
pass	Student-C	O-1.0	Active
fail	Student-C	O-1.0	Active
fail	Student-C	P-1.0	Active
fail	Student-C	P-1.0	Active

我希望我的输出数据框如下：

student	version	pass_count	fail_count	status	total_count (pass+fail)
Student-A	L-1.0	1	2	Active	3
Student-A	M-1.0	1	1	Active	2
Student-B	N-1.0	3	0	Active	3
Student-C	O-1.0	1	1	Active	2
Student-C	P-1.0	0	2	Active	2

我可以使用以下但不是总计数来获得通过和失败计数，请任何人帮助

pd.pivot_table(master_df, index=['status', 'student', 'version'], columns=['result'], aggfunc=len, fill_value=0)

Answer 1

如果要获取总数，可以使用：

df.groupby(["student", "version"])['result'].count().reset_index(name='total_count')

结果：

     student version  total_count
0  Student-A   L-1.0            3
1  Student-A   M-1.0            2
2  Student-B   N-1.0            3
3  Student-C   O-1.0            3
4  Student-C   P-1.0            2

或者，如果您想根据数据透视表获取它，您可以使用：

df2 = pd.pivot_table(df, index=['status', 'student', 'version'], columns=['result'], aggfunc=len, fill_value=0).reset_index().rename_axis(columns=None)

df2['total_count'] = df2['fail'] + df2['pass']

结果：

print(df2)

   status    student version  fail  pass  total_count
0  Active  Student-A   L-1.0     2     1            3
1  Active  Student-A   M-1.0     1     1            2
2  Active  Student-B   N-1.0     0     3            3
3  Active  Student-C   O-1.0     1     2            3
4  Active  Student-C   P-1.0     2     0            2