Question

我想加入表并获得以下输出

表1

TestId1
----------
one
two
three
four
five
six
seven
eight

表2

TestId2
----------
fiftythree
fiftyfour
fiftytwo
fiftyfive
fiftyone

我希望Table3作为输出，包含table1中的所有行和table2中的第一行，直到没有剩余的行，然后它们应该重复开始。

作为替代答案，它们也可以随机分配。

TestId1        TestId2   
----------     ----------
one           fiftythree
two           fiftyfour 
three         fiftytwo  
four          fiftyfive 
five          fiftyone  
six           fiftythree
seven         fiftyfour 
eight         fiftytwo

Answer 1

试试这个：

SELECT t1.name, t2.name FROM
(
    SELECT (ROW_NUMBER() OVER(ORDER BY name)-1)%(SELECT COUNT(*) FROM test2) AS j,* 
    FROM test1
) t1
INNER JOIN 
(
    SELECT ROW_NUMBER() OVER(ORDER BY name)-1 AS j,* 
    FROM test2
) t2 ON t1.j = t2.j
ORDER BY t1.name

详情：

SELECT (ROW_NUMBER() OVER(ORDER BY name)-1) AS j,* 
FROM test1

返回：

0 | one
1 | two
2 | three
3 | four
4 | five
5 | six
6 | seven
7 | eight

此：

SELECT ROW_NUMBER() OVER(ORDER BY name)-1 AS j,* 
FROM test2

返回：

0 | fiftythree
1 | fiftyfour
2 | fiftytwo
3 | fiftyfive
4 | fiftyone

你所要做的就是将最长表的第一列除以（更不知道这个的英文名称）和短句中元素的数量：

SELECT (ROW_NUMBER() OVER(ORDER BY name)-1)%(SELECT COUNT(*) FROM test2) AS j,* 
FROM test1

返回：

0 | one
1 | two
2 | three
3 | four
4 | five
0 | six
1 | seven
2 | eight

现在你要做的就是在第一列上加入两个表。

此解决方案仅使用一个查询，但它假定在table1中，table2中有更多元素。如果你不喜欢这个解决方案，我只是给你很好的基础来编写商店程序。

Answer 2

您的备用解决方案是唯一的解决方案

SELECT
    TestID1,
    TestID2
FROM
    (SELECT COUNT(*) AS Count1 FROM Table1) C1 --one row
    CROSS JOIN
    (SELECT COUNT(*) AS Count2 FROM Table2) C2 --one row
    CROSS JOIN
    (
    SELECT
        ROW_NUMBER() OVER (ORDER BY TestID1) AS Rank1,
        TestID1,
    FROM
        Table1
    ) t1
    JOIN
    (
    SELECT
        ROW_NUMBER() OVER (ORDER BY TestID1) AS Rank2,
        TestID2,
    FROM
        Table2
    ) t2 ON
         t1.Rank1 % CASE WHEN C1.Count1 > C2.Count2 THEN C2.Count2 ELSE 2000000000 END
         =
         t2.Rank2 % CASE WHEN C2.Count2 > C1.Count1 THEN C1.Count1 ELSE 2000000000 END
ORDER BY
    t1.Rank1, t2.Rank2

Answer 3

这里有效，只有一个光标：

if exists(select object_id('tempdb..#TestId1'))
drop table #TestId1

if exists(select object_id('tempdb..#TestId2'))
drop table #TestId2

if exists(select object_id('tempdb..#result'))
drop table #result

create table #TestId1(col_1 varchar(100))

create table #TestId2(col_2 varchar(100))

create table #result (col_1 varchar(100), col_2 varchar(100))

set rowcount 0

insert into #TestId1(col_1 )
select col='one'
union all select col='two'
union all select col='three'
union all select col='four'
union all select col='five'
union all select col='six'
union all select col='seven'
union all select col='eigh'

insert into #TestId2(col_2 )
select col='fiftythree'
union all select col='fiftyfour'
union all select col='fiftytwo'
union all select col='fiftyfive'
union all select col='fiftyone'

DECLARE @sectblcnt int
select @sectblcnt=count(*) from #TestId2

DECLARE @sectableNo int

DECLARE @rowno int

declare @col_1 varchar(100), @col_2 varchar(100)

set @rowno=0

DECLARE curs CURSOR FOR SELECT col_1 FROM #TestId1

OPEN curs

FETCH NEXT FROM curs INTO @col_1

WHILE @@FETCH_STATUS = 0
BEGIN
set @rowno=@rowno+1;

set @sectableNo = @rowno % @sectblcnt
set rowcount @sectableNo

select @col_2=col_2 from #TestId2

insert into #result(col_1, col_2)
    values(@col_1, @col_2)

FETCH NEXT FROM curs 
INTO @col_1
END 

CLOSE curs

DEALLOCATE curs

set rowcount 0

select * from #result

Answer 4

我更喜欢随机过程，所以选择 [伪]随机解决方案 ;-)

这也需要在Table2上“引入”一个行号，但是与Table1的连接是由Table1的某些列上的某些哈希[以Tableol的行数为模]驱动的（不一定是TestId1））。

SELECT T1.TestId1, T2.TestId2 
FROM Table1 T1
JOIN (
  SELECT (ROW_NUMBER() OVER(ORDER BY TestId2) - 1) AS RowNum, TestId2 
  FROM Table2
) T2 ON ABS(HashBytes ('MD5', T1.TestId1) % (SELECT COUNT(*) FROM Table2))
        = T2.RowNum
ORDER BY t1.TestId1

Answer 5

您可能想尝试在ROWNUM上加入。

http://www.adp-gmbh.ch/ora/sql/rownum.html

Answer 6

我认为一个选择的所有解决方案都非常难看。我认为最好使用带有两个游标的存储过程（在每个表上）

连接表（没有子句）

6 个答案: