如何连接表,连接一些数据
问题描述:
我有两个表:
1.索引和索引数量
2.索引和具有指定的boxcodes索引的数量。 Boxcode是一些框,其中包含索引。
如何连接表,连接一些数据
1. input table 1
item_id quantity
1 10
2 15
3 5
1 5
1 5
2 5
3 5
sum:
1 - 20
2 - 20
3 - 10
2. input table 2
item_id quantity boxcode
1 3 abc
2 2 abc
1 8 def
3 10 ghi
1 9 ghi
2 9 def
2 8 ghi !!!!!!!
1 item_id once on 1 boxcode
我希望得到的结果:
从表13. result
item_id quantity boxcodes
1 10 abc/3, def/7
2 15 abc/2, def/9, ghi/4
3 5 ghi/5
1 5 def/1, ghi/4
1 5 ghi/5
2 5 ghi/4 !!!!!!!!
3 5 ghi/5
记录必须以相同的顺序。
我不知道它是如何做到的。
有什么建议吗?
CREATE TABLE #input1
(
rownum int,
item_id int,
quantity int
)
CREATE TABLE #input2
(
item_id int,
quantity int,
boxcode varchar(10)
)
INSERT INTO #input1 VALUES (1,1,10)
INSERT INTO #input1 VALUES (2,2,15)
INSERT INTO #input1 VALUES (3,3,5)
INSERT INTO #input1 VALUES (4,1,5)
INSERT INTO #input1 VALUES (5,1,5)
INSERT INTO #input1 VALUES (6,2,5)
INSERT INTO #input1 VALUES (7,3,5)
INSERT INTO #input2 VALUES (1,3, 'abc')
INSERT INTO #input2 VALUES (2,2, 'abc')
INSERT INTO #input2 VALUES (1,8, 'def')
INSERT INTO #input2 VALUES (3,10, 'ghi')
INSERT INTO #input2 VALUES (1,9, 'ghi')
INSERT INTO #input2 VALUES (2,9, 'def')
INSERT INTO #input2 VALUES (2,8, 'ghi')
select * from #input1
select * from #input2
drop table #input1
drop table #input2
感谢,
答
怪异,但它的工作原理:
;WITH rec1 AS (
SELECT rownum,
item_id,
1 as q,
1 as [Level],
quantity
from #input1
UNION ALL
SELECT r.rownum,
r.item_id,
1,
[Level] + 1,
i.quantity
FROM rec1 r
INNER JOIN #input1 i
ON r.rownum = i.rownum AND r.item_id = i.item_id
WHERE [Level] < i.quantity
), rec2 AS (
SELECT boxcode,
item_id,
1 as q,
1 as [Level],
quantity
from #input2
UNION ALL
SELECT r.boxcode,
r.item_id,
1,
[Level] + 1,
i.quantity
FROM rec2 r
INNER JOIN #input2 i
ON r.boxcode = i.boxcode AND r.item_id = i.item_id
WHERE [Level] < i.quantity
), cte1 AS (
SELECT *,
ROW_NUMBER() OVER (PARTITION BY item_id ORDER BY item_id, rownum) as rn
FROM rec1
), cte2 AS (
SELECT *,
ROW_NUMBER() OVER (PARTITION BY item_id ORDER BY item_id, boxcode) as rn
FROM rec2
), final AS (
SELECT c1.rownum,
c1.item_id,
c1.quantity,
c2.boxcode+'/'+CAST(SUM(c2.q) as nvarchar(10)) as boxcodes
FROM cte1 c1
INNER JOIN cte2 c2
ON c1.item_id = c2.item_id and c1.rn = c2.rn
GROUP BY c1.rownum, c1.item_id, c1.quantity, c2.boxcode
)
SELECT DISTINCT
f.rownum,
f.item_id,
f.quantity,
STUFF((
SELECT ', '+f1.boxcodes
FROM final f1
WHERE f1.rownum = f.rownum
AND f1.item_id = f.item_id
AND f1.quantity = f.quantity
FOR XML PATH('')
),1,2,'') boxcodes
FROM final f
输出的数据集您提供:
rownum item_id quantity boxcodes
1 1 10 abc/3, def/7
2 2 15 abc/2, def/9, ghi/4
3 3 5 ghi/5
4 1 5 def/1, ghi/4
5 1 5 ghi/5
6 2 5 ghi/4
7 3 5 ghi/5
主要想法是在两个表格中传播数量为小零件1
。比添加行号,然后加入并获得结果。
答
一个解决方案(但它完全基于gofr1的答案,说实话!),简化一下,将创建一个数字表,其中包含尽可能多的数字,只要你想。
CREATE TABLE Numbers(Number INT PRIMARY KEY);
INSERT Numbers
SELECT TOP 1000 ROW_NUMBER() OVER (ORDER BY name)
FROM sys.all_columns;
这将避免2递归CTE。
然后,您可以使用相同的逻辑gofr1:
with rec1 AS (
SELECT
ROW_NUMBER() OVER (PARTITION BY item_id ORDER BY item_id, rownum) as rn,
rownum,
item_id,
case when quantity = 0 then 0 else 1 end as q,
quantity
from #input1
join Numbers n on n.Number <= quantity
)
, rec2 AS (
SELECT
ROW_NUMBER() OVER (PARTITION BY item_id ORDER BY item_id, boxcode) as rn,
boxcode,
item_id,
case when quantity = 0 then 0 else 1 end as q,
quantity
from #input2
join Numbers n on n.Number <= quantity
),
final AS (
SELECT c1.rownum,
c1.item_id,
c1.quantity,
c2.boxcode+'/'+CAST(SUM(c2.q) as nvarchar(10)) as boxcodes
FROM rec1 c1
INNER JOIN rec2 c2
ON c1.item_id = c2.item_id and c1.rn = c2.rn
GROUP BY c1.rownum, c1.item_id, c1.quantity, c2.boxcode
),
stuffed as (
SELECT
distinct rownum,
f.item_id,
f.quantity,
STUFF((
SELECT ', '+f1.boxcodes
FROM final f1
WHERE f1.rownum = f.rownum
AND f1.item_id = f.item_id
AND f1.quantity = f.quantity
FOR XML PATH('')
),1,2,'') boxcodes
FROM final f
group by item_id, quantity, boxcodes, rownum)
select *
from stuffed
order by rownum
+0
感谢您编辑我的答案!请注意:OP在他的问题中添加了rownum,请参阅创建脚本的一部分。 – gofr1
+0
@ gofr1你是对的,我删除了我无用的编辑! –
你不必表名或任何所以它很难写答案,但尝试到U SE ROW_NUMBER()OVER(ORDER BY ITEM_ID)上的第一张表 –
如何获得abc/3..def/7 – TheGameiswar
表1中的记录必须是相同的顺序。 我从input_table_1获取第一条记录,并使用item_id1(input_table_2)查找boxcodes。我需要数量:10.我从'abc'框中获得3个,'def'框中获得7个。 3 + 7 = 10 – peter