3秒内剔除同类型交易

Question

例如table如下

类型	时间戳	结果
1	2021-06-2514:21:00	一个
1	2021-06-2514:21:03	B
1	2021-06-2514:21:06	C
1	2021-06-2514:23:00	D
2	2021-06-2514:21:02	C
2	2021-06-2514:21:06	C
2	2021-06-2514:21:09	D
3	2021-06-2514:21:06	E

我想要的结果如下。如果每隔 3 秒或更短时间有一系列相同类型的交易，我将只打印最后一个。

类型	时间戳	结果
1	2021-06-2514:21:06	C
1	2021-06-2514:23:00	D
2	2021-06-2514:21:02	C
2	2021-06-2514:21:09	D
3	2021-06-2514:21:06	E

Answer 1

基本上，您希望相同类型的下一个行提前三秒以上：

select t.*
from (select t.*,
             lead(timestamp) over (partition by type order by timestamp) as next_timestamp
      from t
     ) t
where next_timestamp is null or
      next_timestamp > timestamp + interval '3' second;

Answer 2

使用默认的 LEAD 和 Teradata 的专有 QUALIFY 关键字简化 Gordon 的查询：

select *
from vt
qualify 
   lead(timestamp, 1, date '9999-12-31')
   over (partition by type
         order by timestamp) > timestamp_ + interval '3' second;

3秒内剔除同类型交易

Excluding same type of transactions within 3 seconds

sql

teradata