SQL 查询将列拆分为多行

Question

这很可能以某种方式重复。我查看并尝试了许多找到的答案。我为此苦苦挣扎了好几天，但我无法让它发挥作用，或者弄清楚如何应用我找到的一些示例。

似乎最想将所有列转换为行，这不是我想要的。据我所知，我可能需要的是 unpivot 和/或 UNION？

我正在从 WordPress 表“wp_usermeta”中获取数据。

WordPress 表结构如下：

user_id | meta_key   | meta_value
1       | FirstnameA | Klaus
1       | SecondnameA| Reed
1       | FirstnameB | Herbert
1       | SecondnameB| Hall
2       | FirstnameA | Max
2       | SecondnameA| Musterman
....

我需要什么：

ID | Firstname_A | Lastname_A | Adress_A 

ID | Firstname_B | Lastname_B | Adress_B

两行将属于同一个 ID。

我当前的查询是什么样的：

SELECT user_id,

MAX(CASE WHEN meta_key='firstnameA' THEN meta_value ELSE null END)  FirstnameA,
MAX(CASE WHEN meta_key='lastnameA' THEN meta_value ELSE null END) Lastname_A,
MAX(CASE WHEN meta_key='adressA' THEN meta_value ELSE null END) Adress_A,
MAX(CASE WHEN meta_key='firstnameB' THEN meta_value ELSE null END) FirstnameB,
MAX(CASE WHEN meta_key='lastnameB' THEN meta_value ELSE null END) Lastname_B,
MAX(CASE WHEN meta_key='adressB' THEN meta_value ELSE null END) Adress_B

FROM wp_usermeta
GROUP BY user_id

我的结果是：

ID | Firstname_A | Lastname_A | Adress_A | Firstname_B | Lastname_B | Adress_B

MySQL-版本：5.7

我使用 JOIN for @janh 的方法也可以显示实际的 meta_key：

SET SQL_BIG_SELECTS=1;
select m1.user_id, m1.meta_value as Vorname_M, m2.meta_value as Nachname_M,m3.meta_value as Adresse_M,m4.meta_value as Telefon_M,m5.meta_value as Mail_M,
                   m6.meta_value as Vorname_V, m7.meta_value as Nachname_V,m8.meta_value as Adresse_V,m9.meta_value as Telefon_V,m10.meta_value as Mail_V,
                   m11.meta_value as Kinder, m12.meta_value as Kindergeburtstage,
                   m13.meta_value as Vorname_L, m14.meta_value as Nachname_L,m15.meta_value as Adresse_L,m16.meta_value as Telefon_L,m17.meta_value as Mail_L,
                   m18.meta_value as Arbeitskreise
from wp_usermeta m1
join wp_usermeta m2 on (m1.user_id = m2.user_id and m2.meta_key = 'nachnamemother')
join wp_usermeta m3 on (m2.user_id = m3.user_id and m3.meta_key = 'adress_mother')
join wp_usermeta m4 on (m3.user_id = m4.user_id and m4.meta_key = 'phone_mother')
join wp_usermeta m5 on (m4.user_id = m5.user_id and m5.meta_key = 'mail_mother')
join wp_usermeta m6 on (m5.user_id = m6.user_id and m6.meta_key = 'first_name_father')
join wp_usermeta m7 on (m6.user_id = m7.user_id and m7.meta_key = 'first_name_father_8')
join wp_usermeta m8 on (m7.user_id = m8.user_id and m8.meta_key = 'adress_father')
join wp_usermeta m9 on (m8.user_id = m9.user_id and m9.meta_key = 'phone_mother_16')
join wp_usermeta m10 on (m9.user_id = m10.user_id and m10.meta_key = 'mail_father')

join wp_usermeta m11 on (m10.user_id = m11.user_id and m11.meta_key = 'childs')
join wp_usermeta m12 on (m11.user_id = m12.user_id and m12.meta_key = 'birth_data')

join wp_usermeta m13 on (m12.user_id = m12.user_id and m13.meta_key = 'leben_vorname')
join wp_usermeta m14 on (m13.user_id = m13.user_id and m14.meta_key = 'leben_nachname')
join wp_usermeta m15 on (m14.user_id = m14.user_id and m15.meta_key = 'leben_adress')
join wp_usermeta m16 on (m15.user_id = m15.user_id and m16.meta_key = 'leben_phone')
join wp_usermeta m17 on (m16.user_id = m16.user_id and m17.meta_key = 'leben_mail')

join wp_usermeta m18 on (m17.user_id = m17.user_id and m18.meta_key = 'ak')

where m1.meta_key = 'firstnamemother'

---

什么对我有用：

SELECT user_id,

CASE WHEN meta_key LIKE '%mother' THEN 'mother' 
     WHEN meta_key LIKE '%father' THEN 'father' 
     WHEN meta_key LIKE '%8' THEN 'father'
     END AS AorB,


MAX(CASE WHEN meta_key='firstnamemother' 
           OR meta_key='first_name_father' 
         THEN meta_value END) AS Vorname,
        
MAX(CASE WHEN meta_key='nachnamemother' 
           OR meta_key='first_name_father_8' 
           THEN meta_value END) AS Nachname,
       
 MAX(CASE WHEN meta_key='adress_mother' 
           OR meta_key='adress_father' 
           THEN meta_value END) AS Adresse
       
FROM wp_usermeta
GROUP BY
user_id,
CASE WHEN meta_key LIKE '%mother' THEN 'mother' 
     WHEN meta_key LIKE '%father' THEN 'father' 
     WHEN meta_key LIKE '%8' THEN 'father'
     END

Answer 1

UNION ALL 可以解决问题。

SELECT user_id,
MAX(CASE WHEN meta_key='firstnameA' THEN meta_value ELSE null END)  FirstnameA,
MAX(CASE WHEN meta_key='lastnameA' THEN meta_value ELSE null END) Lastname_A,
MAX(CASE WHEN meta_key='adressA' THEN meta_value ELSE null END) Adress_A
FROM wp_usermeta
GROUP BY user_id
UNION ALL
SELECT user_id,
MAX(CASE WHEN meta_key='firstnameB' THEN meta_value ELSE null END) FirstnameB,
MAX(CASE WHEN meta_key='lastnameB' THEN meta_value ELSE null END) Lastname_B,
MAX(CASE WHEN meta_key='adressB' THEN meta_value ELSE null END) Adress_B
FROM wp_usermeta
GROUP BY user_id

Answer 2

根据您对表格的编辑并希望 A 和 B 作为行而不是单独的列，您可以稍微更改条件聚合并添加一个新列来跟踪记录是 A 还是 B。然后按它和您的聚合进行分组应该给你想要的。

SELECT
    ID
    ,CASE WHEN user_key LIKE '%A' THEN 'A' ELSE 'B' END AS AorB
    ,MAX(CASE WHEN meta_key LIKE 'first%' THEN meta_value END) AS FirstName
    ,MAX(CASE WHEN meta_key LIKE 'last%' OR meta_key LIKE 'sec%' THEN meta_value END) AS LastName
    ,MAX(CASE WHEN meta_key LIKE 'add%' THEN meta_value END) AS Address
FROM
    wp_usermeta
GROUP BY
    user_id
    ,CASE WHEN meta_key LIKE '%A' THEN 'A' ELSE 'B' END

所以您只需要测试firstname 而不是firstnameb，因为如果您测试整个字符串，您将消除A 结果，lastname 和地址也是如此。请注意，在执行case expression 时，如果您不包含ELSE，则与when 语句不匹配的所有内容都将自动为NULL，因此您也可以将其排除。

根据您的评论，您可以在 CASE 表达式中添加额外的 OR 或 WHEN 语句，即使不标准应该只是 x# 的测试用例。加上使用first% 而不是firstname% 将立即占2 个案例，或者您可以考虑f%。最后你可以使用last%和sec%等。

如果您有特定数量的firstnameA,B,C,D。您也可以使用RIGHT(meta_key,1) 而不是case 表达式，只需在group by 中使用相同的东西。如果您有时需要超过 1 个字符而不是其他字符会变得更加困难，但您可以根据 case 表达式进行混合以进行测试，例如CASE WHEN meta_key LIKE '%AB' THEN RIGHT(meta_key,2) ELSE RIGHT(meta_key,1) END

Answer 3

我想连接应该这样做，不是吗？

类似

select ID, pm_firstname.meta_value as firstname, pm_lastname.meta_value lastname, pm_address.meta_value as address FROM wp_posts
    INNER JOIN wp_postmeta pm_firstname ON (wp_posts.ID = pm_firstname.post_id AND pm_firstname.meta_key = 'firstname')
    INNER JOIN wp_postmeta pm_lastname ON (wp_posts.ID = pm_lastname.post_id AND pm_lastname.meta_key = 'lastname')
    INNER JOIN wp_postmeta pm_address ON (wp_posts.ID = pm_address.post_id AND pm_address.meta_key = 'address')
;

应该让你得到很好的组合结果。您需要哪种类型的联接取决于您的数据以及如果任何帖子都不存在字段会发生什么。