如何将可重复分区作为一个又一个分区的新分区?

How to get repeatable partition as new one after another partition?

OVER 子句中的

"Partition by" 将所有值分组为唯一值,就像 "Distinct" 或 "Group by" 一样。

这就是它在我使用 row_number():

的查询中的工作方式
    id st t  row_number
    -------------------
    1  1  1  1
    1  1  2  2
    1  1  3  3
    2  1  3  1
    1  2  4  1
    1  1  10 4

这就是我想要的:

    id st t  uniq_row_number
    ------------------
    1  1  1  1
    1  1  2  2
    1  1  3  3
    2  1  3  1
    1  2  4  1
    1  1  10 1

无论之前是否有new string,每次改变partition后都会读入新的partition。 如果分区重复,那么 uniq_row_number 得到 +1。如果新分区带有新字符串:boom,它会得到 uniq_row_number 1.

我的SQL查询:

    SELECT id, st, t,
    row_number() OVER (PARTITION BY id, st ORDER BY id, st) cat_num,
    min(t) over (PARTITION BY id, st) min_t,
    max(t) over (PARTITION BY id, st) max_t
    FROM tabl ORDER BY t;

SQL 代码在这里:http://sqlfiddle.com/#!18/d4290/2

这称为 "gaps-and-islands" 问题。您需要为每个 "island" 相似值定义一个组。然后你可以使用 row_number().

行号的不同是定义岛的便捷方式:

select t.*,
       row_number() over (partition by id, seqnum_t - seqnum_it
                          order by t
                         ) as uniq_row_number
from (select t.*,
             row_number() over (order by t) as seqnum_t,
             row_number() over (partition by id order by t) as seqnum_it,
      from t
     ) t;

了解其工作原理的最佳方式是查看子查询的结果。您应该能够看到行号的差异如何定义您关心的组。