查询以检索值已更改的行

Question

试图恢复一些非常生疏的 SQL 技能，但已经超出了我的理解范围！我有一个 table（下面的示例），每周我都会上传指示销售机会进展的行。机会具有索引 ID，而某些维度（如“阶段”和“总成本”）可能每周都在变化。每个文件加载还有一个日期格式的 dateLoaded 列。

dateLoaded	OppNumber	Stage	Total Cost
2022-04-25	12345	04	60.00
2022-04-25	23456	01	500.00
2022-04-25	34567	02	225.00
2022-04-25	45678	04	1750.00
2022-04-25	56789	06	50.00
2022-05-01	12345	04	100.00
2022-05-01	23456	01	500.00
2022-05-01	34567	02	275.00
2022-05-01	45678	04	2000.00
2022-05-01	56789	06	50.00
2022-05-07	12345	04	125.00
2022-05-07	23456	02	500.00
2022-05-07	34567	04	275.00
2022-05-07	56789	04	55.00

（注意OppNumber 45678已经在2022-05-07文件中去掉）

我尝试编写的查询将查看 2 个最近的 文件加载，其中“总成本”发生变化，并且 return 都是原始值和新的价值。使用上面的 table 示例，它将 return （按 OppNumber 排序，然后按 dateLoaded ASC 排序）：

dateLoaded	OppNumber	Stage	Total Cost
2022-05-01	12345	04	100.00
2022-05-07	12345	04	125.00
2022-05-01	56789	06	50.00
2022-05-07	56789	06	55.00

理想情况下，我也很想知道 2 次加载前而不是最近的记录存在于何处（在本例中，它将 return 2022-05-01 中 45678 的记录文件）。

我最终不得不将这些数据转储到 Excel 并创建公式，但我确信有办法为此编写查询。

提前感谢您的建议！

Answer 1

如果你有 mysql 8 或以上，你可以使用 window 函数

示例 1

第一个 cte 选择那些在最近的负载中有一个 oppnumber 的那些并且 dense_ranks 按日期并且滞后以获得之前的价格之后它非常简单

WITH CTE AS 
(
SELECT * ,DENSE_RANK() OVER (ORDER BY DATELOADED) DR,
         LAG(TOTAL_COST) OVER (PARTITION BY OPPNUMBER ORDER BY DATELOADED) PREV
FROM T
WHERE EXISTS (SELECT OPPNUMBER FROM T T1 WHERE T1.OPPNUMBER = T.OPPNUMBER AND T1.DATELOADED = (sELECT MAX(DATELOADED) FROM T) )
ORDER BY OPPNUMBER,DATELOADED DESC
),
CTE1 AS
(SELECT MAX(DR) MAXDR FROM CTE
) 
SELECT CTE.*
FROM CTE 
JOIN CTE1 ON (CTE.DR = CTE1.MAXDR)
WHERE TOTAL_COST <> PREV
UNION ALL
SELECT CTE.*
FROM CTE 
JOIN CTE1 ON (CTE.DR = CTE1.MAXDR - 1)
JOIN (SELECT CTE.*
FROM CTE 
JOIN CTE1 ON (CTE.DR = CTE1.MAXDR)
WHERE TOTAL_COST <> PREV) S ON S.OPPNUMBER = CTE.OPPNUMBER
ORDER BY OPPNUMBER, DATELOADED ASC;

+------------+-----------+-------+------------+----+--------+
| dateLoaded | OppNumber | Stage | Total_Cost | DR | PREV   |
+------------+-----------+-------+------------+----+--------+
| 2022-05-01 |     12345 |     4 |     100.00 |  2 |  60.00 |
| 2022-05-07 |     12345 |     4 |     125.00 |  3 | 100.00 |
| 2022-05-01 |     56789 |     6 |      50.00 |  2 |  50.00 |
| 2022-05-07 |     56789 |     4 |      55.00 |  3 |  50.00 |
+------------+-----------+-------+------------+----+--------+
4 rows in set (0.028 sec)

示例 2

与示例 1 类似，除了特别关注最近的变化而不是最近的负载

WITH CTE AS 
(
SELECT * ,row_number() OVER (PARTITION BY OPPNUMBER ORDER BY DATELOADED) rn,
         LAG(TOTAL_COST) OVER (PARTITION BY OPPNUMBER ORDER BY DATELOADED) PREV
FROM T
ORDER BY OPPNUMBER,DATELOADED DESC
),
cte1 as
(
Select oppnumber opn, max(rn) maxrn from cte where total_cost <> prev group by oppnumber
)
select * from cte join cte1 on cte1.opn = cte.oppnumber and cte1.maxrn = cte.rn
union all
select * from cte join cte1 on cte1.opn = cte.oppnumber and cte1.maxrn - 1 = cte.rn
order by oppnumber , dateloaded desc

+------------+-----------+-------+------------+----+---------+-------+-------+
| dateLoaded | OppNumber | Stage | Total_Cost | rn | PREV    | opn   | maxrn |
+------------+-----------+-------+------------+----+---------+-------+-------+
| 2022-05-07 |     12345 |     4 |     125.00 |  3 |  100.00 | 12345 |     3 |
| 2022-05-01 |     12345 |     4 |     100.00 |  2 |   60.00 | 12345 |     3 |
| 2022-05-01 |     34567 |     2 |     275.00 |  2 |  225.00 | 34567 |     2 |
| 2022-04-25 |     34567 |     2 |     225.00 |  1 |    NULL | 34567 |     2 |
| 2022-05-01 |     45678 |     4 |    2000.00 |  2 | 1750.00 | 45678 |     2 |
| 2022-04-25 |     45678 |     4 |    1750.00 |  1 |    NULL | 45678 |     2 |
| 2022-05-07 |     56789 |     4 |      55.00 |  3 |   50.00 | 56789 |     3 |
| 2022-05-01 |     56789 |     6 |      50.00 |  2 |   50.00 | 56789 |     3 |
+------------+-----------+-------+------------+----+---------+-------+-------+
8 rows in set (0.004 sec)

查询以检索值已更改的行

Query to retrieve rows where a value has changed

mysql