问题描述
我有根据资源 ID 为每个用户分配项目的表格。我们正在创建一些需要通过从 Staffed 值中减去 Demand 值来获得超额分配值的仓库。
需求值是唯一的，基于 Number,ResourceId 的组合。
+-----------------------------------------------------------------------------------------------------+
| Number | ResourceId | Demand01 | Demand02 | Staffed01 | Staffed02 | AssociateName | GroupName |
+-----------------------------------------------------------------------------------------------------+
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.30 | 0.30 | Raja | RR/A |
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.70 | 0.70 | Praveen | RR/A |
| RR-00000001 | 1020688 | 0.00 | 0.00 | 0.12 | 1.00 | Bala | RR/A |
| RR-00000002 | 1025136 | 0.00 | 0.00 | 0.00 | 0.00 | Naveen | RR/B |
| RR-00000003 | 1020258 | 0.01 | 0.01 | 0.90 | 0.90 | Kumar | RR/C |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | Arun | RR/D |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.70 | 0.70 | BBB | RR/E |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | CCC | RR/E |
+-----------------------------------------------------------------------------------------------------+
预期输出：
+-----------------------------------------------------------------------------------------------------------------------------------------+
| Number | ResourceId | Demand01 | Demand02 | Staffed01 | Staffed02 | OverStaffed01 | OverStaffed02 | AssociateName | GroupName |
+-----------------------------------------------------------------------------------------------------------------------------------------+
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.30 | 0.30 | 0.00 | 0.00 | Raja | RR/A |
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.70 | 0.70 | -0.60 | -0.60 | Praveen | RR/A |
| RR-00000001 | 1020688 | 0.00 | 0.00 | 0.12 | 1.00 | -0.12 | -1.00 | Bala | RR/A |
| RR-00000002 | 1025136 | 0.00 | 0.10 | 0.00 | 0.20 | 0.00 | -0.20 | Naveen | RR/B |
| RR-00000003 | 1020258 | 0.01 | 0.01 | 0.90 | 0.90 | -0.89 | -0.89 | Kumar | RR/C |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | -0.29 | -0.29 | Arun | RR/D |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.40 | 0.40 | -0.40 | -0.40 | BBB | RR/E |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | -0.30 | -0.30 | CCC | RR/E |
+-----------------------------------------------------------------------------------------------------------------------------------------+
我必须推导出 OverStaffed01 = Demand01 - Staffed01 但有一次约束，如果我们有重复
Number 和 ResourceId，我们必须根据前一行调整需求值。
例如：
Number:RR-00000002 and ResourceId:1019096 的组合有三行，所以我们不能直接计算值。
对于第一行，(OverStaffed01 = Demand01 - Staffed01) 0.40 - 0.30 是 大于 0 表示人员不足，我们认为是 0。
对于第二行 (OverStaffed01 = Demand01 - Staffed01) 0.10 - 0.70 （因为我们已经从第一行减少了 0.30，所以我们有 Demand01 为 0.10)，而 OverStaffed01 为 -0.60
Number:RR-00000001 and ResourceId:1019094 的组合有两行，所以我们不能直接计算值。
对于第一行，(OverStaffed01 = Demand01 - Staffed01) 0.01 - 0.30 是 -0.29。
对于第二行 (OverStaffed01 = Demand01 - Staffed01) 0.00 - 0.40 （因为我们已经从第一行减少了 0.01，所以我们有 Demand01 作为 0) 和 OverStaffed01 是 -0.40
对于第三行 (OverStaffed01 = Demand01 - Staffed01) 0.00 - 0.30 （因为我们已经从第一行减少了 0.01，所以我们有 Demand01 为 0)，而 OverStaffed01 为 -0.30
对于其他行，我们有唯一的Number和ResourceId，所以我们可以直接推算OverStaffed01
到目前为止尝试过的查询：
CREATE TABLE table1
(
Number varchar(100),ResourceId varchar(100),Demand01 decimal(18,2),Demand02 decimal(18,Staffed01 decimal(18,Staffed02 decimal(18,AssociateName varchar(100),GroupName varchar(100)
)
INSERT INTO table1 VALUES('RR-00000001','1019094','0.40','0.30','Raja','RR/A')
INSERT INTO table1 VALUES('RR-00000001','0.70','Praveen','1020688','0.00','0.12','1.00','Bala','RR/A')
INSERT INTO table1 VALUES('RR-00000002','1025136','0.10','0.20','Naveen','RR/B')
INSERT INTO table1 VALUES('RR-00000003','1020258','0.01','0.90','Kumar','RR/C')
INSERT INTO table1 VALUES('RR-00000002','1019096','Arun','RR/D')
INSERT INTO table1 VALUES('RR-00000002','BBB','RR/E')
INSERT INTO table1 VALUES('RR-00000002','CCC','RR/E')
SELECT Number,ResourceId,Demand01,Staffed01,AssociateName,GroupName,ROW_NUMBER() OVER (PARTITION BY Number,ResourceId ORDER BY (Demand01 - Staffed01) DESC,(Demand02 - Staffed02) DESC) RN,Demand01- SUM (Staffed01) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL) ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS OverStaffed01,Demand02- SUM (Staffed02) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL) ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) AS OverStaffed02
FROM table1
更新
我已经发布了我最近更新的查询，它以某种方式给出了正确的结果，但仍然不是预期的结果。
解决方法
架构：
CREATE TABLE table1 (
Number varchar(100),ResourceId varchar(100),Demand01 decimal(18,2),Demand02 decimal(18,Staffed01 decimal(18,Staffed02 decimal(18,AssociateName varchar(100),GroupName varchar(100)
)
INSERT INTO table1 VALUES('RR-00000001','1019094','0.40','0.30','Raja','RR/A')
INSERT INTO table1 VALUES('RR-00000001','0.70','Praveen','1020688','0.00','0.12','1.00','Bala','RR/A')
INSERT INTO table1 VALUES('RR-00000002','1025136','0.10','0.20','Naveen','RR/B')
INSERT INTO table1 VALUES('RR-00000003','1020258','0.01','0.90','Kumar','RR/C')
INSERT INTO table1 VALUES('RR-00000002','1019096','Arun','RR/D')
INSERT INTO table1 VALUES('RR-00000002','BBB','RR/E')
INSERT INTO table1 VALUES('RR-00000002','CCC','RR/E')
查询：
with cte as (
SELECT Number,ResourceId,Demand01,Demand02,Staffed01,Staffed02,AssociateName,GroupName,RN,coalesce(sum(staffed01)over (partition by number,resourceid order by rn rows between unbounded preceding and 1 preceding),0) PreviousStaffed01,coalesce(sum(staffed02)over (partition by number,0) PreviousStaffed02
FROM
(
SELECT Number,ROW_NUMBER() OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL)) RN
FROM table1
) X
)
select Number,(case when ((case when (demand01-previousstaffed01) <0 then 0 else (demand01-previousstaffed01)end) -staffed01)<0
then ((case when (demand01-previousstaffed01) <0 then 0 else (demand01-previousstaffed01)end) -staffed01) else 0 end) OverStaffed01,(case when ((case when (demand02-previousstaffed02) <0 then 0 else (demand02-previousstaffed02)end) -staffed02)<0
then ((case when (demand02-previousstaffed02) <0 then 0 else (demand02-previousstaffed02)end) -staffed02) else 0 end) OverStaffed02,GroupName
from cte
输出：
| 数字 | ResourceId | Demand01 | Demand02 | Staffed01 | Staffed02 | OverStaffed01 | OverStaffed02 | AssociateName | GroupName |
|---|---|---|---|---|---|---|---|---|---|
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.30 | 0.30 | 0.00 | 0.00 | 王爷 | RR/A |
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.70 | 0.70 | -0.60 | -0.60 | Praveen | RR/A |
| RR-00000001 | 1020688 | 0.00 | 0.00 | 0.12 | 1.00 | -0.12 | -1.00 | 巴拉 | RR/A |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | -0.29 | -0.29 | 阿伦 | RR/D |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.40 | 0.40 | -0.40 | -0.40 | BBB | RR/E |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | -0.30 | -0.30 | CCC | RR/E |
| RR-00000002 | 1025136 | 0.00 | 0.10 | 0.00 | 0.20 | 0.00 | -0.10 | Naveen | RR/B |
| RR-00000003 | 1020258 | 0.01 | 0.01 | 0.90 | 0.90 | -0.89 | -0.89 | 库马尔 | RR/C |
dbfiddle here
,感谢您的回答Kazi Mohammad Ali Nur
我尝试编写查询 ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW，但查询有点复杂。
SELECT Number,CASE WHEN RN = 1 THEN OverStaffed01
ELSE CASE WHEN SIGN(LAG(OverStaffed01,1,0) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL))) = 0 THEN OverStaffed01
ELSE 0 - Staffed01
END
END OverStaffed01,CASE WHEN RN = 1 THEN OverStaffed02
ELSE CASE WHEN SIGN(LAG(OverStaffed02,ResourceId ORDER BY (SELECT NULL))) = 0 THEN OverStaffed02
ELSE 0 - Staffed02
END
END OverStaffed02
FROM
(
SELECT Number,ResourceId ORDER BY (Demand01 - Staffed01) DESC,(Demand02 - Staffed02) DESC) RN,CASE WHEN Demand01 - SUM (Staffed01) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL) ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) > 0 THEN 0
ELSE Demand01 - SUM (Staffed01) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL) ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
END AS OverStaffed01,CASE WHEN Demand02 - SUM (Staffed02) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL) ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW) > 0 THEN 0
ELSE Demand02 - SUM (Staffed02) OVER (PARTITION BY Number,ResourceId ORDER BY (SELECT NULL) ROWS BETWEEN UNBOUNDED PRECEDING AND CURRENT ROW)
END AS OverStaffed02
FROM table1
)Tmp
+-----------------------------------------------------------------------------------------------------------------------------------------+
| Number | ResourceId | Demand01 | Demand02 | Staffed01 | Staffed02 | OverStaffed01 | OverStaffed02 | AssociateName | GroupName |
+-----------------------------------------------------------------------------------------------------------------------------------------+
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.30 | 0.30 | 0.00 | 0.00 | Raja | RR/A |
| RR-00000001 | 1019094 | 0.40 | 0.40 | 0.70 | 0.70 | -0.60 | -0.60 | Praveen | RR/A |
| RR-00000001 | 1020688 | 0.00 | 0.00 | 0.12 | 1.00 | -0.12 | -1.00 | Bala | RR/A |
| RR-00000002 | 1025136 | 0.00 | 0.10 | 0.00 | 0.20 | 0.00 | -0.20 | Naveen | RR/B |
| RR-00000003 | 1020258 | 0.01 | 0.01 | 0.90 | 0.90 | -0.89 | -0.89 | Kumar | RR/C |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | -0.29 | -0.29 | Arun | RR/D |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.40 | 0.40 | -0.40 | -0.40 | BBB | RR/E |
| RR-00000002 | 1019096 | 0.01 | 0.01 | 0.30 | 0.30 | -0.30 | -0.30 | CCC | RR/E |
+-----------------------------------------------------------------------------------------------------------------------------------------+