如果完全重复 - Python，则用另一列的子集为每个唯一帐户填充新列 True/False

Question

我需要使用 True / False 值创建新列（未评级重复？）。如果子集（开始、结束、数量、Sig、费率）对于每个唯一日期范围（开始、结束）都是完全重复的，则将 True 分配给特定值的行（收费？== 'N'）。

原始 df:

 Account      Start     End         Qty   Sig   Rate Charged?
1234567890  2021-08-18  2021-09-17  1174    A   ABC     N   
1234567890  2021-08-18  2021-09-17  0       B   ABC     N   
1234567890  2021-08-18  2021-09-17  0       C   ABC     N   
1234567890  2021-08-18  2021-09-17  224     D   ABC     N   
1234567890  2021-08-18  2021-09-17  1398   TOT  ABC     N   
1234567890  2021-08-18  2021-09-17  2024    A   ABC     Y   
1234567890  2021-08-18  2021-09-17  0       B   ABC     Y   
1234567890  2021-08-18  2021-09-17  0       C   ABC     Y   
1234567890  2021-08-18  2021-09-17  414     D   ABC     Y   
1234567890  2021-08-18  2021-09-17  2438   TOT  ABC     Y   
1234567890  2021-09-17  2021-10-16  548     A   ABC     N   
1234567890  2021-09-17  2021-10-16  586     B   ABC     N   
1234567890  2021-09-17  2021-10-16  113     C   ABC     N   
1234567890  2021-09-17  2021-10-16  105     D   ABC     N   
1234567890  2021-09-17  2021-10-16  1352   TOT  ABC     N   
1234567890  2021-09-17  2021-10-16  548     A   ABC     Y   
1234567890  2021-09-17  2021-10-16  586     B   ABC     Y   
1234567890  2021-09-17  2021-10-16  113     C   ABC     Y   
1234567890  2021-09-17  2021-10-16  105     D   ABC     Y   
1234567890  2021-09-17  2021-10-16  1352   TOT  ABC     Y

输出应该是：

 Account      Start     End         Qty   Sig   Rate Charged? Unr Dupl ?
1234567890  2021-08-18  2021-09-17  1174    A   ABC     N        False
1234567890  2021-08-18  2021-09-17  0       B   ABC     N        False
1234567890  2021-08-18  2021-09-17  0       C   ABC     N        False
1234567890  2021-08-18  2021-09-17  224     D   ABC     N        False
1234567890  2021-08-18  2021-09-17  1398   TOT  ABC     N        False
1234567890  2021-08-18  2021-09-17  2024    A   ABC     Y        False
1234567890  2021-08-18  2021-09-17  0       B   ABC     Y        False
1234567890  2021-08-18  2021-09-17  0       C   ABC     Y        False
1234567890  2021-08-18  2021-09-17  414     D   ABC     Y        False
1234567890  2021-08-18  2021-09-17  2438   TOT  ABC     Y        False
1234567890  2021-09-17  2021-10-16  548     A   ABC     N        True
1234567890  2021-09-17  2021-10-16  586     B   ABC     N        True
1234567890  2021-09-17  2021-10-16  113     C   ABC     N        True
1234567890  2021-09-17  2021-10-16  105     D   ABC     N        True
1234567890  2021-09-17  2021-10-16  1352   TOT  ABC     N        True
1234567890  2021-09-17  2021-10-16  548     A   ABC     Y        False
1234567890  2021-09-17  2021-10-16  586     B   ABC     Y        False
1234567890  2021-09-17  2021-10-16  113     C   ABC     Y        False
1234567890  2021-09-17  2021-10-16  105     D   ABC     Y        False
1234567890  2021-09-17  2021-10-16  1352   TOT  ABC     Y        False

我的代码，但它不能按我想要的方式工作：

duplicate_read = []

for i in df["Account"].unique():
    
    for ind in df[df["Account"] == i].index:
    
        duplicate_read.append(df[df["Account"] == i].duplicated(subset = ["Start", "End", "Qty", "Sig", "Rate"], keep = 'last')[ind])

df["Unrated Duplicate ?"] = duplicate_read

df

谢谢。

Answer 1

尝试：

df["Duplicate"] = df.duplicated(["Start", "End", "Qty", "Sig", "Rate"], keep=False)
df["Unr Dupl?"] = df.groupby(["Start","End"])["Duplicate"].transform("all")&df["Charged?"].eq("N")
df = df.drop("Duplicate", axis=1)

>>> df
       Account       Start         End   Qty  Sig Rate Charged?  Unr Dupl?
0   1234567890  2021-08-18  2021-09-17  1174    A  ABC        N      False
1   1234567890  2021-08-18  2021-09-17     0    B  ABC        N      False
2   1234567890  2021-08-18  2021-09-17     0    C  ABC        N      False
3   1234567890  2021-08-18  2021-09-17   224    D  ABC        N      False
4   1234567890  2021-08-18  2021-09-17  1398  TOT  ABC        N      False
5   1234567890  2021-08-18  2021-09-17  2024    A  ABC        Y      False
6   1234567890  2021-08-18  2021-09-17     0    B  ABC        Y      False
7   1234567890  2021-08-18  2021-09-17     0    C  ABC        Y      False
8   1234567890  2021-08-18  2021-09-17   414    D  ABC        Y      False
9   1234567890  2021-08-18  2021-09-17  2438  TOT  ABC        Y      False
10  1234567890  2021-09-17  2021-10-16   548    A  ABC        N       True
11  1234567890  2021-09-17  2021-10-16   586    B  ABC        N       True
12  1234567890  2021-09-17  2021-10-16   113    C  ABC        N       True
13  1234567890  2021-09-17  2021-10-16   105    D  ABC        N       True
14  1234567890  2021-09-17  2021-10-16  1352  TOT  ABC        N       True
15  1234567890  2021-09-17  2021-10-16   548    A  ABC        Y      False
16  1234567890  2021-09-17  2021-10-16   586    B  ABC        Y      False
17  1234567890  2021-09-17  2021-10-16   113    C  ABC        Y      False
18  1234567890  2021-09-17  2021-10-16   105    D  ABC        Y      False
19  1234567890  2021-09-17  2021-10-16  1352  TOT  ABC        Y      False

如果完全重复 - Python，则用另一列的子集为每个唯一帐户填充新列 True/False

Fill new column with True/False for each unique account with subset of another columns if exact duplicate - Python

python

boolean

duplicates

pandas