有没有办法使用 rank 函数将 table 中的值展平？

Question

我在 SQL 中有一个 table 看起来像这样 -

ITEM	ACTIVITY_ID	ACTIVITY_TYPE	ACTIVITY_DATE
Item 1	Activity A	Call	Jan - 1 - 2022
Item 1	Activity B	Mail	Jan - 10 - 2022
Item 1	Activity C	Print	Jan - 12 - 2022

同样，有数千个项目，每个项目可以有一个或多个活动（最多 5 个）。我想运行一个 SQL 查询来展平所有记录的项目级别的数据，所需的输出是这样的 -

ITEM	ACTIVITY 1	ACTIVITY 2	ACTIVITY 3	ACTIVITY 4	ACTIVITY 5	ACTIVITY 1 DATE	ACTIVITY 2 DATE	ACTIVITY 3 DATE	ACTIVITY 4 DATE	ACTIVITY 5 DATE
Item 1	Call	Mail	Print			Jan - 1 - 2022	Jan - 10 - 2022	Jan - 12 - 2022

activity 列 (1-5) 是根据 activity 日期的升序填充的。

有办法实现吗？此外，我可以在 Python 中导入原始数据，如果有使用 Pandas.

的优雅方法，我也可以在那里进行转换

请注意，列值仍保留为列值，它与 pandas 中的逆透视操作不同。我在 pandas 中看到了关于 unpivoting 的答案，但无法使用那里的答案解决这个特定问题

提前致谢，

Answer 1

模式：

WITH cte AS (
    SELECT *, ROW_NUMBER() OVER (PARTITION BY item ORDER BY activity_date) rn
    FROM source_table
)
SELECT t1.item,
       t1.activity_type activity_1,
--     ...
       t5.activity_type activity_5,
       t1.activity_date date_1,
--     ...
       t5.activity_date date_5
FROM cte t1
LEFT JOIN cte t2 ON t1.item = t2.item AND t2.rn = 2
LEFT JOIN cte t3 ON t1.item = t3.item AND t3.rn = 3
LEFT JOIN cte t4 ON t1.item = t4.item AND t4.rn = 4
LEFT JOIN cte t5 ON t1.item = t5.item AND t5.rn = 5
WHERE t1.rn = 1

PS。 activity_date 列中的数据格式似乎是 non-standard，可能需要转换为 DATE 数据类型。

Answer 2

您正在寻找 PIVOT，而不是 UNPIVOT。

但在您的情况下，NPath 也适用：

SELECT * 
FROM NPath
       ( ON (
               SELECT ITEM, ACTIVITY_TYPE, ACTIVITY_DATE
               FROM tab
             )
         PARTITION BY ITEM                   -- group by column
         ORDER BY ACTIVITY_DATE              -- order within list
         USING                                      
           MODE (NonOverlapping)             -- required syntax 
           Symbols (True AS T)               -- every row
           Pattern ('T*')                    --   is aggregated
           RESULT(First (item OF T) AS item  -- group by column
                 ,First (ACTIVITY_TYPE OF T) AS activity_1_type
                 ,NTH (ACTIVITY_TYPE,2 OF T) AS activity_2_type
                 ,NTH (ACTIVITY_TYPE,3 OF T) AS activity_3_type
                 ,NTH (ACTIVITY_TYPE,4 OF T) AS activity_4_type
                 ,NTH (ACTIVITY_TYPE,5 OF T) AS activity_5_type
                 ,First (ACTIVITY_DATE OF T) AS activity_1_date
                 ,NTH (ACTIVITY_DATE,2 OF T) AS activity_2_date
                 ,NTH (ACTIVITY_DATE,3 OF T) AS activity_3_date
                 ,NTH (ACTIVITY_DATE,4 OF T) AS activity_4_date
                 ,NTH (ACTIVITY_DATE,5 OF T) AS activity_5_date
                 ,Count(* OF T)
                 )
        )
;

有没有办法使用 rank 函数将 table 中的值展平？

Is there a way to flatten values in a table using rank function?

python

mysql

sql

pandas

teradata