Question

我有以下数据。

a_id    state   date

1   A   7/5/19

1   A   7/6/19

1   M   7/7/19

1   M   7/8/19

1   A   7/9/19

1   A   7/10/19

1   A   7/11/19

1   M   7/12/19

我想使用SQL为每个组按顺序到达它们。

例如，A有2个组，因为它们没有以相同的顺序到达。对于M同样请参见下面的预期输出。

a_id    state   date    result col

1   A   7/5/19  1

1   A   7/6/19  1

1   M   7/7/19  2

1   M   7/8/19  2

1   A   7/9/19  3

1   A   7/10/19 3

1   A   7/11/19 3

1   M   7/12/19 4

Answer 1

这是一个孤岛问题。您可以使用lag()和累计金额：

select t.*,
       sum(case when state = prev_state then 0 else 1 end) over (partition by a_od order by date) as result
from (select t.*,
             lag(state) over (partition by a_id order by date) as prev_state
      from t
     ) t;

SQL为每个组生成通用序列

1 个答案: