尝试从联接表中选择一些行,其中有多个相关行。
这是原始查询的非常简化的版本,因此请不要对常规表结构提出任何建议。
表A
id, title
------------
1, housea
2, houseb
3, housec
4, housed
表B
id, cid, attrib, val
--------------------
1, 1, bathrooms, 2
2, 1, bedrooms, 1
3, 2, bathrooms, 0
4, 1, pools, 1
5, 2, bedrooms, 1
6, 2, pools, 1
7, 3, bathrooms, 1
8, 4, bathrooms, 1
9, 4, bedrooms, 1
选择所有至少具有一间浴室和一间卧室的对象。
因此,只有这两个应出现:
2, housea
4, housed
这不起作用:
SELECT a.id, title
FROM tablea a
LEFT JOIN tableb b ON b.cid = a.id
WHERE (b.attrib = "bathrooms" AND b.val > 0) AND (b.attrib = "bedrooms" AND b.val > 0)
也不是:
SELECT a.id, title
FROM tablea a
LEFT JOIN tableb b1 ON b1.cid = a.id AND (b1.attrib = "bathrooms" AND b1.val > 0)
LEFT JOIN tableb b2 ON b2.cid = a.id AND (b2.attrib = "bedrooms" AND b2.val > 0)
感谢您的建议!
答案 0 :(得分:1)
您的第二个版本基本上是正确的,除了您需要内部联接:
SELECT a.id, a.title
FROM tablea a JOIN
tableb b1
ON b1.cid = a.id AND (b1.attrib = 'bathrooms' AND b1.val > 0) JOIN
tableb b2
ON b2.cid = a.id AND (b2.attrib = 'bedrooms' AND b2.val > 0) ;
您所有的过滤都在ON
子句中,因此实际上没有任何行被过滤掉。如果查看b1
或b2
中的列,则会看到过滤条件。
答案 1 :(得分:0)
尝试如下运行Correlated子查询:
select a.id, a.title from tablea a where 2 >=
(select count(b.cid) from tableb b where b.cid = a.id and b.attrib in ('bathrooms' , 'bedrooms') and b.val > 0 group by b.cid)