有条件 summing/subtracting 数量
Conditionally summing/subtracting quantities
我有两个 table（PS_BU_ITEMS_INV A 和 PS_TRANSACTION_INV B）用于库存项目管理。 PS_BU_ITEMS_INV A 包含一个字段，其中包含项目 A.QTY_ONHAND 的当前现有数量。 table PS_TRANSACTION_INV B 是一项交易 table，它列出了减少或增加 [=19] 中的 A.QTY_ONHAND 值的所有交易（装运、消耗等） =] table 来自名为 B.QTY_BASE 的字段。
B.QTY_BASE 列中存储的值都是正数，因此我必须查看另一个名为 B.TRANSACTION_GROUP 的列来确定 B.QTY_BASE 中的数量是否应该添加或减去当前 A.QTY_ONHAND 值。
下面是每个 tables
的示例数据
PS_BU_ITEMS_INV:
SELECT BUSINESS_UNIT, INV_ITEM_ID, QTY_ONHAND, DT_TIMESTAMP
FROM PS_BU_ITEMS_INV
WHERE INV_ITEM_ID = '1'
AND BUSINESS_UNIT = '11MMS'
结果：
BUSINESS_UNIT INV_ITEM_ID QTY_ONHAND DT_TIMESTAMP
11MMS 1 16.0000 2018-09-11 08:12:46.827
PS_TRANSATION_INV:
SELECT BUSINESS_UNIT, INV_ITEM_ID, TRANSACTION_DATE,
TRANSACTION_GROUP, QTY_BASE, DT_TIMESTAMP
FROM PS_TRANSACTION_INV
WHERE INV_ITEM_ID = '1'
AND BUSINESS_UNIT = '11MMS'
ORDER BY DT_TIMESTAMP DESC
结果：
假设我想确定 9 月 4 日的 QTY_ONHAND 值；我将从 PS_BU_ITEMS_INV table 中获取当前 QTY_ONHAND 值（在本例中为 16.0 并添加（总和）（来自 PS_TRANSACTION_INV）相关的 QTY_BASE 值使用 Transaction_Group(s) '030', and '036'（这些是物品消耗）并使用 TRANSATION_GROUP = '020'
减去数量
所以我应该有截至 2018 年 9 月 4 日的 ON_QTY 计算：16 + (SUM OF -->(14+2+2+5+2+1+2+10- 24+10+10-24) = 26
到目前为止，我已经将以下查询写到 SUM 数量（对于今天（当前日期）和硬编码日期（9/4/ 18 在这个例子中），但我正在努力如何添加某种类型的 CASE WHEN 或条件 IF 语句，如果 TRANSACTION_GROUP 是 030 或 036 然后将值加在一起，如果 TRANSACTION_GROUP 是 020，则减去这些值。由于缺少条件减法，查询目前无法从 2018 年 9 月 4 日得到正确的 ON_HAND_QTY。
SELECT A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + (SELECT SUM(BB.QTY_BASE)
FROM PS_TRANSACTION_INV BB WHERE BB.TRANSACTION_DATE
BETWEEN '2018-09-04' AND GETDATE() AND BB.INV_ITEM_ID
= '1'
AND BB.BUSINESS_UNIT = '11MMS'
AND BB.BUSINESS_UNIT = B.BUSINESS_UNIT AND BB.INV_ITEM_ID
= B.INV_ITEM_ID)) AS QTY_ONHAND_AS_OF_DT
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
我想要一些类似这样的效果：
SELECT
A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + (CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') THEN
SUM(B.QTY_BASE)
WHEN B.TRANSACTION_GROUP = '020' THEN
SUM(-B.QTY_BASE) END AS ON_HANDS_AS_OF_DATE ) )
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
虽然这个查询不起作用。感谢任何反馈。
编辑：
这是我编辑的 SQL，但它 return 是错误的有条件计算的现有价值。我什至在日期范围的 CASE WHEN 中添加了附加条件，但它仍然没有 return 正确的值 (26)。
SELECT
A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + SUM(CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN (B.QTY_BASE)
WHEN B.TRANSACTION_GROUP = '020' --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN (-B.QTY_BASE) END ) )
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
GROUP BY A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP, DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
编辑 2：
我很好奇为什么这个查询得到我正确的数量计数 (26)（为了简化，我硬编码了 A.QTY_ONHAND 中的值）
SELECT (32 + SUM(CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE() THEN
THEN (B.QTY_BASE)
WHEN B.TRANSACTION_GROUP = '020' --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN (-B.QTY_BASE) END ) )
FROM PS_TRANSACTION_INV B
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
然而下面的查询（使用相同的条件逻辑）得到了不同的（不正确的）数量？
SELECT
A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + SUM(CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE() THEN
THEN B.QTY_BASE
WHEN B.TRANSACTION_GROUP = '020' --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN -B.QTY_BASE END ) )
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
GROUP BY A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP, DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
我不会调试您的确切查询。不过，您可以在下面看到如何有条件地对值求和并将其减去当前值，这样您就可以回到过去并找到原始数量：
假设您的数据是：
create table ps_bu_items_inv (
business_unit varchar(20),
inv_item_id int,
qty_onhand decimal(12, 8),
dt_timestamp datetime
);
insert into ps_bu_items_inv (business_unit, inv_item_id, qty_onhand, dt_timestamp)
values ('11MMS', 1, 16.0, '2018-09-11 08:12:46.827');
create table ps_transaction_inv (
business_unit varchar(20),
inv_item_id int,
transaction_date date,
transaction_group varchar(6),
qty_base decimal(12, 8),
dt_timestamp datetime
);
insert into ps_transaction_inv (business_unit, inv_item_id, transaction_date, transaction_group, qty_base, dt_timestamp)
values ('11MMS', 1, '2018-09-11', '036', 14, '2018-09-11 12:34:56');
insert into ps_transaction_inv (business_unit, inv_item_id, transaction_date, transaction_group, qty_base, dt_timestamp)
values ('11MMS', 1, '2018-09-11', '020', 24, '2018-09-11 12:34:56');
insert into ps_transaction_inv (business_unit, inv_item_id, transaction_date, transaction_group, qty_base, dt_timestamp)
values ('11MMS', 1, '2018-09-11', '030', 6, '2018-09-11 12:34:56');
查询应类似于：
select a.*, b.*, (a.qty_onhand - b.change) as original_qty
from ps_bu_items_inv a
join (
select
inv_item_id,
sum( -- here's the conditional sum
case when transaction_group <> '020' then -qty_base else qty_base end
) as change
from ps_transaction_inv
where transaction_date between '2018-09-01' and '2018-09-11'
group by inv_item_id
) b on a.inv_item_id = b.inv_item_id
结果：
business_unit inv_item_id qty_onhand dt_timestamp inv_item_id change original_qty
------------- ----------- ---------- ------------ ----------- ------ ------------
11MMS 1 16 2018-09-11 1 4 12
如您所见，原始数量显示为 12。即当前 16 减去变化 4。在这种情况下，"changes" 计算为：
-14 +24 -6 = +4
我有两个 table（PS_BU_ITEMS_INV A 和 PS_TRANSACTION_INV B）用于库存项目管理。 PS_BU_ITEMS_INV A 包含一个字段，其中包含项目 A.QTY_ONHAND 的当前现有数量。 table PS_TRANSACTION_INV B 是一项交易 table，它列出了减少或增加 [=19] 中的 A.QTY_ONHAND 值的所有交易（装运、消耗等） =] table 来自名为 B.QTY_BASE 的字段。
B.QTY_BASE 列中存储的值都是正数，因此我必须查看另一个名为 B.TRANSACTION_GROUP 的列来确定 B.QTY_BASE 中的数量是否应该添加或减去当前 A.QTY_ONHAND 值。
下面是每个 tables
的示例数据PS_BU_ITEMS_INV:
SELECT BUSINESS_UNIT, INV_ITEM_ID, QTY_ONHAND, DT_TIMESTAMP
FROM PS_BU_ITEMS_INV
WHERE INV_ITEM_ID = '1'
AND BUSINESS_UNIT = '11MMS'
结果：
BUSINESS_UNIT INV_ITEM_ID QTY_ONHAND DT_TIMESTAMP
11MMS 1 16.0000 2018-09-11 08:12:46.827
PS_TRANSATION_INV:
SELECT BUSINESS_UNIT, INV_ITEM_ID, TRANSACTION_DATE,
TRANSACTION_GROUP, QTY_BASE, DT_TIMESTAMP
FROM PS_TRANSACTION_INV
WHERE INV_ITEM_ID = '1'
AND BUSINESS_UNIT = '11MMS'
ORDER BY DT_TIMESTAMP DESC
结果：
假设我想确定 9 月 4 日的 QTY_ONHAND 值；我将从 PS_BU_ITEMS_INV table 中获取当前 QTY_ONHAND 值（在本例中为 16.0 并添加（总和）（来自 PS_TRANSACTION_INV）相关的 QTY_BASE 值使用 Transaction_Group(s) '030', and '036'（这些是物品消耗）并使用 TRANSATION_GROUP = '020'
所以我应该有截至 2018 年 9 月 4 日的 ON_QTY 计算：16 + (SUM OF -->(14+2+2+5+2+1+2+10- 24+10+10-24) = 26
到目前为止，我已经将以下查询写到 SUM 数量（对于今天（当前日期）和硬编码日期（9/4/ 18 在这个例子中），但我正在努力如何添加某种类型的 CASE WHEN 或条件 IF 语句，如果 TRANSACTION_GROUP 是 030 或 036 然后将值加在一起，如果 TRANSACTION_GROUP 是 020，则减去这些值。由于缺少条件减法，查询目前无法从 2018 年 9 月 4 日得到正确的 ON_HAND_QTY。
SELECT A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + (SELECT SUM(BB.QTY_BASE)
FROM PS_TRANSACTION_INV BB WHERE BB.TRANSACTION_DATE
BETWEEN '2018-09-04' AND GETDATE() AND BB.INV_ITEM_ID
= '1'
AND BB.BUSINESS_UNIT = '11MMS'
AND BB.BUSINESS_UNIT = B.BUSINESS_UNIT AND BB.INV_ITEM_ID
= B.INV_ITEM_ID)) AS QTY_ONHAND_AS_OF_DT
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
我想要一些类似这样的效果：
SELECT
A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + (CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') THEN
SUM(B.QTY_BASE)
WHEN B.TRANSACTION_GROUP = '020' THEN
SUM(-B.QTY_BASE) END AS ON_HANDS_AS_OF_DATE ) )
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
虽然这个查询不起作用。感谢任何反馈。
编辑：
这是我编辑的 SQL，但它 return 是错误的有条件计算的现有价值。我什至在日期范围的 CASE WHEN 中添加了附加条件，但它仍然没有 return 正确的值 (26)。
SELECT
A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + SUM(CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN (B.QTY_BASE)
WHEN B.TRANSACTION_GROUP = '020' --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN (-B.QTY_BASE) END ) )
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
GROUP BY A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP, DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
编辑 2：
我很好奇为什么这个查询得到我正确的数量计数 (26)（为了简化，我硬编码了 A.QTY_ONHAND 中的值）
SELECT (32 + SUM(CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE() THEN
THEN (B.QTY_BASE)
WHEN B.TRANSACTION_GROUP = '020' --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN (-B.QTY_BASE) END ) )
FROM PS_TRANSACTION_INV B
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
然而下面的查询（使用相同的条件逻辑）得到了不同的（不正确的）数量？
SELECT
A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP,
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0) AS TRANSACTION_DT,
(A.QTY_ONHAND + SUM(CASE WHEN B.TRANSACTION_GROUP IN ('030', '036') --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE() THEN
THEN B.QTY_BASE
WHEN B.TRANSACTION_GROUP = '020' --AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
THEN -B.QTY_BASE END ) )
FROM PS_BU_ITEMS_INV A
LEFT OUTER JOIN PS_TRANSACTION_INV B ON B.BUSINESS_UNIT
= A.BUSINESS_UNIT
AND B.INV_ITEM_ID = A.INV_ITEM_ID
AND DATEADD(MINUTE, DATEDIFF(MINUTE, 0, A.DT_TIMESTAMP), 0) =
DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
WHERE B.INV_ITEM_ID = '1'
AND B.BUSINESS_UNIT = '11MMS'
AND B.TRANSACTION_DATE BETWEEN '2018-09-04' AND GETDATE()
GROUP BY A.INV_ITEM_ID, A.BUSINESS_UNIT, A.QTY_ONHAND, A.DT_TIMESTAMP, DATEADD(MINUTE, DATEDIFF(MINUTE, 0, B.DT_TIMESTAMP), 0)
我不会调试您的确切查询。不过，您可以在下面看到如何有条件地对值求和并将其减去当前值，这样您就可以回到过去并找到原始数量：
假设您的数据是：
create table ps_bu_items_inv (
business_unit varchar(20),
inv_item_id int,
qty_onhand decimal(12, 8),
dt_timestamp datetime
);
insert into ps_bu_items_inv (business_unit, inv_item_id, qty_onhand, dt_timestamp)
values ('11MMS', 1, 16.0, '2018-09-11 08:12:46.827');
create table ps_transaction_inv (
business_unit varchar(20),
inv_item_id int,
transaction_date date,
transaction_group varchar(6),
qty_base decimal(12, 8),
dt_timestamp datetime
);
insert into ps_transaction_inv (business_unit, inv_item_id, transaction_date, transaction_group, qty_base, dt_timestamp)
values ('11MMS', 1, '2018-09-11', '036', 14, '2018-09-11 12:34:56');
insert into ps_transaction_inv (business_unit, inv_item_id, transaction_date, transaction_group, qty_base, dt_timestamp)
values ('11MMS', 1, '2018-09-11', '020', 24, '2018-09-11 12:34:56');
insert into ps_transaction_inv (business_unit, inv_item_id, transaction_date, transaction_group, qty_base, dt_timestamp)
values ('11MMS', 1, '2018-09-11', '030', 6, '2018-09-11 12:34:56');
查询应类似于：
select a.*, b.*, (a.qty_onhand - b.change) as original_qty
from ps_bu_items_inv a
join (
select
inv_item_id,
sum( -- here's the conditional sum
case when transaction_group <> '020' then -qty_base else qty_base end
) as change
from ps_transaction_inv
where transaction_date between '2018-09-01' and '2018-09-11'
group by inv_item_id
) b on a.inv_item_id = b.inv_item_id
结果：
business_unit inv_item_id qty_onhand dt_timestamp inv_item_id change original_qty
------------- ----------- ---------- ------------ ----------- ------ ------------
11MMS 1 16 2018-09-11 1 4 12
如您所见，原始数量显示为 12。即当前 16 减去变化 4。在这种情况下，"changes" 计算为：
-14 +24 -6 = +4