MySQL子查询WHERE过滤掉了太多

Question

更短的版本:

SQLfiddle

检查输出列lives_together.我希望它包含与当前成员居住在同一地址但不是他/她的父母或子女的成员的连续(逗号分隔)ID.

所以在第一行,约翰,我想要玛丽的ID(仅限).
不是约瑟夫,因为他是约翰的孩子,而不是维多利亚,因为她住在另一个地址(她也是他的孩子).

现在我什么都没得到,不知怎的,这是因为查询中的这些行(在第三个子查询中):

mem.id <> m3.parent1 AND
mem.id <> m3.parent2 AND
mem.parent1 <> m3.id AND
mem.parent2 <> m3.id AND

他们应该过滤掉所有父母和孩子(他们这样做),但由于某种原因,他们也过滤掉了其他成员(上例中的玛丽).
只有其中一行就足以过滤掉Mary.
另一个例子是Mats和Gabriella,他们应该得到彼此的ID,但他们没有.

(所有人living_together(包括孩子)的一栏是不够的,因为我想living_together在孩子面前打印这个人.这个问题是我刚才问过的另一个问题的发展.我想要的结果有一个很大的不同是它永远不应该将具有不同姓氏的人分组.)

为什么会这样？

更长的版本

我有一张桌子members.有点简化它看起来像这样(你可以在下面的SQLfiddle中完整地看到它):

+-------+------------+------------+------------+---------+----------+----------+-----------+-----------+---------------+
| id    | first_name | last_name  | birthdate  | parent1 | parent2  | address  | phones    | mobiles   | emails        |
+-------+------------+------------+------------+---------+----------+----------+-----------+-----------+---------------+
| 2490  | Mary       | Johansen   | 1964-01-24 | NULL    | NULL     | Street 1 | 1111      | 9999,8888 | mary@test.com |
| 2491  | John       | Johansen   | 1968-01-21 | NULL    | NULL     | Street 1 | 1111,3333 | 7777      | john@test.com |
| 2422  | Brad       | Johansen   | 1983-01-07 | 2491    | 2490     | Street 1 | 2222,3333 | 6666      | brad@test.com |
| 2493  | Victoria   | Andersen   | 1982-01-14 | 2490    | 2491     | Av. 2    | 4444      | 5555      | vic@text.com  |
+-------+------------+------------+------------+---------+----------+----------+-----------+-----------+---------------+

玛丽和约翰结婚了,有两个孩子; 约瑟夫和维多利亚.维多利亚搬出去了.

我想打印一个地址列表,该列表将居住在同一地址的人分组.所以玛丽,约翰和约瑟夫应该分组,但维多利亚应该单独检索.儿童应该在配偶(以及居住在该地址的任何其他成员)之后打印,这是单独检索它们的原因.

以下是我的数据库的真实外观以及我的整个查询:

http://sqlfiddle.com/#!2/1f87c4/2 这与实际数据类似,该表包含大约400行.

所以我想要的是lives_together输出列,包含group_concat匹配成员的inated ID.所以我应该id在Mary的lives_together专栏中找到John ,反之亦然.而不是孩子们的ids.

我期待类似于以下内容.
请注意,尽管在字母表中较早的字母上有一个较低的id和一个名字,但是在John之后显示Brad,这是因为他是该家庭中的孩子.孩子应该按出生日期排序(你可以看到该栏目中的那一栏
也注意到他们的电话号码与他们的姓氏一起添加,但他们的手机号码(和电子邮件)都带有他们的名字

+----------------------------------------------------
| Andersen         Av 2       4444
|     Victoria                5555              vic@test.com
+----------------------------------------------------
| Johansen         Street 1   1111,2222,3333
|     John                    7777              john@test.com
|     Mary                    9999,8888         mary@test.com
|     Brad                    6666              brad@test.com
+----------------------------------------------------

但是,这不是我对下面的查询的期望,而是我在PHP处理之后的目标.

这是我想要的查询:

 +------+----------+-----------+------------+----------+-----------+-----------+---------------+-----------+-----------+----------------+
 | id   | lname    | fname     | birthdate  | address  | phones    | mobiles   | emails        | parents   | children  | lives_together |
 +------+----------+-----------+------------+----------+-----------+-----------+---------------+-----------+-----------+----------------+
 | 2490 | Johansen | Mary      | 1964-01-24 | Street 1 | 1111      | 9999,8888 | mary@test.com | NULL      | 2424      | 2491           |
 +------+----------+-----------+------------+----------+-----------+-----------+---------------+-----------+-----------+----------------+
 | 2491 | Johansen | John      | 1968-01-21 | Street 1 | 1111,3333 | 7777      | john@test.com | NULL      | 2424      | 2490           |
 +------+----------+-----------+------------+----------+-----------+-----------+---------------+-----------+-----------+----------------+
 | 2422 | Johansen | Brad      | 1983-01-07 | Street 1 | 2222,3333 | 6666      | brad@test.com | 2490,2491 | NULL      | NULL           |
 +------+----------+-----------+------------+----------+-----------+-----------+---------------+-----------+-----------+----------------+
 | 2493 | Andersen | Victoria  | 1982-01-14 | Av. 2    | 4444      | 5555      | vic@test.com  | NULL      | NULL      | NULL           |
 +------+----------+-----------+------------+----------+-----------+-----------+---------------+-----------+-----------+----------------+

这是我的查询(也有点简化):

SELECT 
    mem.id as id,
    mem.last_name as lname,
    mem.first_name as fname,
    mem.birthdate as birthdate,
    mem.address as address,
    mem.phones as phones,
    mem.mobiles as mobiles

    (SELECT
        GROUP_CONCAT(m1.id)
        FROM 
            members m1
        WHERE 
            (mem.parent1 = m1.id OR mem.parent2 = m1.id) AND
            LOWER(REPLACE(mem.last_name, ' ', '')) = LOWER(REPLACE(m1.last_name, ' ', '')) AND
            LOWER(REPLACE(mem.address, ' ', '')) = LOWER(REPLACE(m1.address, ' ', '')) AND
            mem.id <> m1.id
    ) as parents,

    (SELECT 
        GROUP_CONCAT(m2.id)
        FROM 
            members m2
        WHERE 
            (mem.id = m2.parent1 OR mem.id = m2.parent2) AND
            LOWER(REPLACE(mem.last_name, ' ', '')) = LOWER(REPLACE(m2.last_name, ' ', '')) AND
            LOWER(REPLACE(mem.address, ' ', '')) = LOWER(REPLACE(m2.address, ' ', '')) AND
            mem.id <> m2.id 
    ) as children,

    (SELECT 
        GROUP_CONCAT(m3.id)
        FROM 
            members m3
        WHERE 
            mem.id <> m3.parent1 AND
            mem.id <> m3.parent2 AND
            mem.parent1 <> m3.id AND
            mem.parent2 <> m3.id AND
            LOWER(REPLACE(mem.last_name, ' ', '')) = LOWER(REPLACE(m3.last_name, ' ', '')) AND
            LOWER(REPLACE(mem.address, ' ', '')) = LOWER(REPLACE(m3.address, ' ', '')) AND
            mem.id <> m3.id 
    ) as lives_together 

FROM 
    members mem 

ORDER BY
    mem.last_name ASC,
    mem.first_name ASC 

当查询在Mary时,并通过第三个子查询(lives_together)它应该得到约翰,但不是约瑟夫或维多利亚.不是约瑟夫因为mem.parent2 <> m3.id而不是维多利亚因为LOWER(REPLACE(mem.address, ' ', '')) = LOWER(REPLACE(m3.address, ' ', '')).这是有效的,但由于某种原因,约翰也没有,我得到了NULL.

如果删除这部分,我会得到约翰,约瑟夫和维多利亚:

mem.id <> m3.parent1 AND
mem.id <> m3.parent2 AND
mem.parent1 <> m3.id AND
mem.parent2 <> m3.id AND

但是当它在那里我没有得到它们.我不明白为什么那部分过滤掉约翰.(请注意,这些行根本没有简化.)只有其中一行用于过滤掉约翰,但按照预期为Josef工作.

我的代码有问题吗？

Answer 1

我不确定这是否是您正在寻找的，我认为您需要每个地址都有一条记录，并且每个地址列出居住在该地址的其他任何人。

SELECT distinct 
    mem.id as id,
    mem.last_name as lname ,
    mem.first_name as fname,
    mem.birth_date as birthdate,
    mem.phone_number as phone,
    mem.mobile_number as mobile,
    mem.email_address as email,
    mem.co_address as co,
    mem.street_address as street,
    mem.postal_address as postal,
    mem.country as country,
    (SELECT 
        GROUP_CONCAT(m3.id)
        FROM 
            members m3
        WHERE 
            LOWER(REPLACE(mem.last_name, ' ', '')) = LOWER(REPLACE(m3.last_name, ' ', '')) AND
            LOWER(REPLACE(mem.street_address, ' ', '')) = LOWER(REPLACE(m3.street_address, ' ', '')) AND
            LOWER(REPLACE(mem.postal_address, ' ', '')) = LOWER(REPLACE(m3.postal_address, ' ', '')) AND
            LOWER(REPLACE(mem.country, ' ', '')) = LOWER(REPLACE(m3.country, ' ', '')) AND
            mem.id <> m3.id 
            AND m3.member_type = 1
    ) as lives_together
FROM 
    members mem

WHERE
    member_type = 1
group by co_address,street_address,postal_address,country
ORDER BY
    mem.last_name ASC,
    mem.first_name ASC,
    mem.birth_date DESC