Teradata SQL - 在一列中找到最后一个匹配项，然后从下面的行中找到 return 值

Question

我有一个如下所示的 table：我已经按时间、ID 订购了 table。我想为每个 ID 找到最后一个事件 44 并获取它下面的值。

Table:

ID    Time    Event
A   00:10:00    11
A   00:10:11    44
A   00:10:13    22
A   00:10:00    11
A   00:10:11    44
A   00:10:13    33
A   00:10:13    22
B   01:10:00    ?
B   01:10:11    44
B   01:10:13    11
B   01:10:00    44
B   01:10:11    ?
B   01:10:13    22
...   ...         ...

想要的结果：

ID    Time    Event
A   00:10:13    33
B   01:10:13    22
... ... ...

已经尝试了不同的方法，但仍在努力...任何帮助将不胜感激。非常感谢！

Answer 1

假设您有一列实际上表示您确实想要下一行的顺序和，这是一种方法：

select t.*
from (select t.*,
             max(case when event = 44 then next_ordercol end) over (partition by id order by ordercol) as event44_next
      from (select t.*,
                   lead(ordercol) over (partition by id order by ordercol) as next_ordercol
            from t
           ) t
     ) t
where ordercol = event44_next;

我总是忘记 Teradata 没有实现 lead() 和 lag()。我真的不明白。这是一种方法：

      from (select t.*,
                   max(ordercol) over (partition by id order by ordercol rows between 1 following and 1 following) as next_ordercol
            from t
           ) t

Answer 2

这将导致解释中的单个 STAT FUNCTION 步骤：

SELECT *
FROM tab
WHERE event IS NOT NULL
QUALIFY
    -- previous event was 44 
        Min(event) -- LEAD
        Over (PARTITION BY id
              ORDER BY "time" DESC 
              ROWS BETWEEN 1 Following AND 1 Following) = 44 
    AND
    -- only return rows when there was no previous 44 event
        Count(CASE WHEN event = '44' THEN 1 end)
        Over (PARTITION BY id
              ORDER BY "time" DESC ROWS Unbounded Preceding) = 0

例如

ID       TIME   Event    Lead  Count
 A   00:11:14      22      33      0
 A   00:11:13      33      44      0
 A   00:11:11      44      11      1
 A   00:11:00      11      22      1
 A   00:10:13      22      44      1
 A   00:10:11      44      11      2

如果你运行 TD16+ 终于支持 LAG/LEAD 而不是 MIN 重写：

Lead(event)
Over (PARTITION BY id
      ORDER BY "time" DESC)

可以在没有 WHERE 事件 IS NOT NULL 的情况下使用 LAST_VALUE（或 LEAD）加上 IGNORE NULLS 选项进行重写，但这将导致解释中的两个 STAT FUNCTION 步骤。

Teradata SQL - 在一列中找到最后一个匹配项，然后从下面的行中找到 return 值

Teradata SQL - find the last match in a column then return value from row below

sql

teradata