一对多数据表联合搜索的问题

[离子漂浮物]

数据表：School[学校]
字段：
scId 编号
scName 名称

数据表：Classes[班级]
字段：
clId 编号
clSchoolId 所属学校编号
clName 班级名称

数据表：Students[学生]
字段：
stId 编号
stClassesId 所属班级编号
stName 姓名

要求显示出 学生姓名或班级名字或学校名称里包含有“文”字的学校中按学校编号倒序排序的第5到第10个学校。

数据记录大至有50万+以上，所以对SQL优化的要求很高，作起来头痛死，有知道的兄弟帮帮忙吧。

[离子漂浮物]

[Quote=引用 3 楼 llsen 的回复:]
写的有点乱
时间不够了

就是加上你的条件
然后出去前五名后的前五名，就是6-10位，不过这样效率可能有点低
你这么大的数据量
[/Quote]

呵，老实说，没看懂你写的SQL句..

[yangqidong]

会啊。

[xudongdong1990]

挺复杂的啊，帮顶了

[pl1069]

请用sql2000写

[离子漂浮物]

[Quote=引用 8 楼 yangqidong 的回复:]
会啊。
[/Quote]

那..不是会很慢很慢？？..

[wuyq11]

create view v_schoole
as
select school .* from school left join Classes on school.scId=clSchoolId left join Students on stClassesId =clId

SELECT * FROM (SELECT TOP 5 * FROM (SELECT TOP 10 * from v_schoole ORDER BY scid DESC) AS T ORDER BY scid ASC) AS T1 ORDER BY scid DESC

[llsen]

[Quote=引用 12 楼 wuyq11 的回复:]
create view v_schoole
as
select school .* from school left join Classes on school.scId=clSchoolId left join Students on stClassesId =clId

SELECT * FROM (SELECT TOP 5 * FROM (SELECT TOP 10 * from v_schoole ORDER BY scid DESC) AS T ORDER BY scid ASC) AS T1 ORDER BY scid DESC
[/Quote]

这个顺序反了，是按从10到6的顺序

[sayoath]

你不用ROW_NUMBER TOP只能说是最好的了。以前的分页还是用TOP呢


把语句改成存储过程。可以解决超时问题。不会那么慢的。

[jidianxueyuan]

select scId,sName from
(
select sc.scId,sc.sName,select row_number()over (order by productname) as rownumber
from School sc inner join Classes cl on sc.scId=cl.clSchoolId
inner join Students st on st.stClassesId=cl.clId
where sc.scName like '%文%' or cl.clName like '%文%' or st.stName like '%文%'
order by scId desc
) as Result
where rownumber between 5 and 10

[离子漂浮物]

[Quote=引用 16 楼 jidianxueyuan 的回复:]
select scId,sName from
(
select sc.scId,sc.sName,select row_number()over (order by productname) as rownumber
from School sc inner join Classes cl on sc.scId=cl.clSchoolId
inner join Students st on st.stClassesId=cl.clId
where sc.scName like '%文%' or cl.clName like '%文%' or st.stName like '%文%'
order by scId desc
) as Result
where rownumber between 5 and 10
[/Quote]

row_number()是MSSQL 2005的函数，我的数据库是MSSQL2000。

[离子漂浮物]

[Quote=引用 12 楼 wuyq11 的回复:]
create view v_schoole
as
select school .* from school left join Classes on school.scId=clSchoolId left join Students on stClassesId =clId

SELECT * FROM (SELECT TOP 5 * FROM (SELECT TOP 10 * from v_schoole ORDER BY scid DESC) AS T ORDER BY scid ASC) AS T1 ORDER BY scid DESC
[/Quote]


一个学校里有多个班级，一个班级里有多个学生。
如果A学校有十个符合条件的记录的话，这句SQL返回的学校将会是十行A学校的记录，而不是十个学校记录
如果符合条件的记录有十万以上的话，估计会运行超时吧...

[陌上花花]

帮顶了

[离子漂浮物]

[Quote=引用 19 楼 zzxap 的回复:]


SQL code
用临时表+2次top
SELECT a.*,t.* into #tablea from Students a left join
(select * from Classes b left join School c on b.clSchoolId=c.scId )t
on a.stClassesId=t.clId
where charindex('文',a.stName)>0 and charindex('文',t.clName)>0 and charindex('文',t.scName)>0

select top 5 from (select top 10 * from #table order by clSchoolId asc )t order by clSchoolId desc
[/Quote]


如果第五到第六个的学校是同一个学校，那不出来的都是五个一样的学校了？

[离子漂浮物]

如果第五到第十个的学校是同一个学校，那不出来的都是五个一样的学校了？

[沉序员]

顶一下，

[zzxap]

[code=SQL]
用临时表+2次top
SELECT a.*,t.* into #tablea from Students a left join
(select * from Classes b left join School c on b.clSchoolId=c.scId )t
on a.stClassesId=t.clId
where charindex('文',a.stName)>0 and charindex('文',t.clName)>0 and charindex('文',t.scName)>0

select top 5 from (select top 10 * from #table order by clSchoolId asc )t order by clSchoolId desc

[/CODE]

[makun0624]

学习`关注`UP

[redhat2]

关注

.



人力资源

[springbell]

顶个

[mengxj85]

[Quote=引用 17 楼 oyiboy 的回复:]
引用 16 楼 jidianxueyuan 的回复:
select scId,sName from
(
select sc.scId,sc.sName,select row_number()over (order by productname) as rownumber
from School sc inner join Classes cl on sc.scId=cl.clSchoolId
inner join Students st on st.stClassesId=cl.clId
where sc.scName like '%文%' or cl.clName like '%文%' or st.stName like '%文%'
order by scId desc
) as Result
where rownumber bet…
[/Quote]
利用临时表，加一列自动编号的列就可以了