将运行索引添加到按用户id分区的pandas中

Question

给定此数据集

CUSTOMER_ID,ORDER_AT
A,2020-11-11 23:30:13
A,2020-11-11 23:32:53
A,2020-11-11 23:34:44
A,2020-11-11 23:35:55
B,2020-11-11 23:37:54
B,2020-11-11 23:39:23
C,2020-11-09 23:59:46
C,2020-11-10 0:03:04
C,2020-11-10 0:05:35
C,2020-11-10 0:19:40
C,2020-11-11 2:48:17
C,2020-11-11 2:49:06
C,2020-11-11 2:50:39
C,2020-11-11 2:51:57
D,2020-11-14 1:12:52
D,2020-11-14 1:13:14
D,2020-11-14 16:56:18

如何创建由CUSTOMER_ID分区的事务id的正在运行的二级索引？

我想要的输出是

CUSTOMER_ID,CUSTOMER_TRANSACTION_ID,ORDER_AT
A,0,2020-11-11 23:30:13
A,1,2020-11-11 23:32:53
A,2,2020-11-11 23:34:44
A,3,2020-11-11 23:35:55
B,0,2020-11-11 23:37:54
B,1,2020-11-11 23:39:23
C,0,2020-11-09 23:59:46
C,1,2020-11-10 0:03:04
C,2,2020-11-10 0:05:35
C,3,2020-11-10 0:19:40
C,4,2020-11-11 2:48:17
C,5,2020-11-11 2:49:06
C,6,2020-11-11 2:50:39
C,7,2020-11-11 2:51:57
D,0,2020-11-14 1:12:52
D,1,2020-11-14 1:13:14
D,2,2020-11-14 16:56:18

Answer 1

试试Groupby和cumcount()。

df['CUSTOMER_TRANSACTION_ID']=df.groupby('CUSTOMER_ID').cumcount()