在 pandas 中分組和聚合

Question

    code_presentation   code_module score   id_student  id_assessment   date_submitted
0   2013J               AAA         78.0    11391        1752           18
1   2013J               AAA         70.0    11391        1800           22
2   2013J               AAA         72.0    31604        1752           17
3   2013J               AAA         69.0    31604        1800           26
.....

我需要計算提交的天數以及如何正確分組以獲得以下結果：

id_student  id_assessment date_submitted
11391       1752          1
            1800          1
31604       1752          1
            1800          1

... ETC

我嘗試：

analasys_grouped = analasys.groupby ( 'id_student', as_index = False)\
.agg({'id_assessment':'count', 'date_submitted': 'count'})
analasys_grouped 

但它工作不正常

Answer 1

如果我對您的理解正確，您想對按id_assessment分組的id_student應用value_counts() 。 嘗試：

assessment_count_per_student = df.groupby('id_student')['id_assessment'].value_counts()

print(assessment_count_per_student)

id_student  id_assessment
11391       1752             1
            1800             1
31604       1752             1
            1800             1
Name: id_assessment, dtype: int64

Answer 2

您需要將id_assessment傳遞到groupby語句中。

df.groupby(['id_student', 'id_assessment'])['date_submitted'].count()


id_student  id_assessment
11391       1752             1
            1800             1
31604       1752             1
            1800             1

---

在您的嘗試中，您僅按id_student分組，然后計算提交的評估和日期。