mysql 连接两个表，并获取匹配记录的字段名称，其值为非零值答案

【问题标题】：mysql join two table, and get field name of the matched record which has nonzero valuemysql 连接两个表，并获取匹配记录的字段名称，其值为非零值
【发布时间】：2017-01-25 16:17:44
【问题描述】：

我有两张如下表

mysql> show tables;
+-------------------+
| Tables_in_testdbs |
+-------------------+
| dts               |
| ref               |
+-------------------+
2 rows in set (0.00 sec)

各表内容如下

mysql> select * from ref;
+----+------+------+
| Id | key1 | key2 |
+----+------+------+
|  1 |    1 |    1 |
|  2 |    1 |    2 |
|  3 |    2 |    2 |
|  4 |    3 |    1 |
|  5 |    3 |    2 |
|  6 |    3 |    3 |
+----+------+------+
6 rows in set (0.00 sec)

mysql> select * from dts;
+----+------+------+--------+------+------+------+------+------+
| Id | key1 | key2 | serial | pr1  | pr2  | pr3  | pr4  | pr5  |
+----+------+------+--------+------+------+------+------+------+
|  1 |    1 |    1 |      1 |    0 |    0 |    1 |    0 |    2 |
|  2 |    1 |    1 |      2 |    0 |    0 |    0 |    0 |    0 |
|  3 |    1 |    1 |      3 |    0 |    0 |    0 |    1 |    0 |
|  4 |    1 |    1 |      4 |    1 |    0 |    1 |    1 |    3 |
|  5 |    1 |    2 |      5 |    0 |    0 |    0 |    2 |    5 |
|  6 |    1 |    2 |      6 |    0 |    0 |    0 |    0 |    1 |
|  7 |    1 |    2 |      7 |    0 |    1 |    0 |    0 |    0 |
|  8 |    2 |    2 |      1 |    1 |    1 |    1 |    1 |    2 |
|  9 |    2 |    2 |      2 |    0 |    0 |    0 |    0 |    0 |
| 10 |    3 |    2 |      3 |    0 |    0 |    0 |    0 |    0 |
| 11 |    3 |    3 |      1 |    1 |    1 |    0 |    0 |    1 |
| 12 |    3 |    3 |      5 |    0 |    0 |    1 |    1 |    0 |
+----+------+------+--------+------+------+------+------+------+
12 rows in set (0.00 sec)

这是我尝试加入两个表的原因

mysql> select distinct
    ->        i.key1,
    ->        i.key2 
    -> from 
    ->        ref i, 
    ->        dts d 
    -> where 
    ->        i.key1=d.key1 and 
    ->        i.key2=d.key2 ;
+------+------+
| key1 | key2 |
+------+------+
|    1 |    1 |
|    1 |    2 |
|    2 |    2 |
|    3 |    2 |
|    3 |    3 |
+------+------+
5 rows in set (0.00 sec)

我期待低于o/p，真的不知道如何得到它

key1       key2 fields_non_zero
1           1   pr1,pr3,pr4,pr5
1           2   pr2,pr4,pr5
2           2   pr1,pr2,pr3,pr4,pr5
3           2
3           3   pr1,pr2,pr3,pr4,pr5

我想使用以下条件进行检查，例如，让 key1=1 和 key2=1 匹配两个表

加入两个表

检查dts的字段（pr1-pr5）中是否有匹配的非零数据

如果发现用逗号连接字段名，

假设如果所有字段都不为零，只需连接字段并停止进一步加入相同的 key1，key2 因为如果找到所有（保存执行时间），转到下一个key1，key2

+----+------+------+
| Id | key1 | key2 |
+----+------+------+
|  1 |    1 |    1 |              <- for ref table key1,key2 following rows matches 

| Id | key1 | key2 | serial | pr1  | pr2  | pr3  | pr4  | pr5  |  nonzero_fields 
+----+------+------+--------+------+------+------+------+------+
|  1 |    1 |    1 |      1 |    0 |    0 |    1 |    0 |    2 |  = pr3,pr5
|  2 |    1 |    1 |      2 |    0 |    0 |    0 |    0 |    0 |  =  
|  3 |    1 |    1 |      3 |    0 |    0 |    0 |    1 |    0 |  = pr4
|  4 |    1 |    1 |      4 |    1 |    0 |    1 |    1 |    3 |  = pr1,pr3,pr4,p45

  So distinct of below are

         = pr3,pr5
         = 
         = pr4
         = pr1,pr3,pr4,p45 

 key1   key2 fields_non_zero
 1       1   pr1,pr3,pr4,pr5

如果我没有像下面这样的订单，我不介意至少

 key1   key2 fields_non_zero
 1       1   pr3,pr5,pr4,pr1

表格结构如下

DROP TABLE IF EXISTS `dts`;
CREATE TABLE `dts` (
  `Id` int(11) NOT NULL AUTO_INCREMENT,
  `key1` int(11) DEFAULT '-99',
  `key2` int(11) DEFAULT '-99',
  `serial` int(11) DEFAULT '-99',
  `pr1` int(11) DEFAULT '-99',
  `pr2` int(11) DEFAULT '-99',
  `pr3` int(11) DEFAULT '-99',
  `pr4` int(11) DEFAULT '-99',
  `pr5` int(11) DEFAULT '-99',
  PRIMARY KEY (`Id`),
  KEY `main` (`key1`,`key2`,`serial`)
) ENGINE=MyISAM AUTO_INCREMENT=13 DEFAULT CHARSET=latin1;


LOCK TABLES `dts` WRITE;
INSERT INTO `dts` VALUES (1,1,1,1,0,0,1,0,2),(2,1,1,2,0,0,0,0,0),(3,1,1,3,0,0,0,1,0),(4,1,1,4,1,0,1,1,3),(5,1,2,5,0,0,0,2,5),(6,1,2,6,0,0,0,0,1),(7,1,2,7,0,1,0,0,0),(8,2,2,1,1,1,1,1,2),(9,2,2,2,0,0,0,0,0),(10,3,2,3,0,0,0,0,0),(11,3,3,1,1,1,0,0,1),(12,3,3,5,0,0,1,1,0);
UNLOCK TABLES;



DROP TABLE IF EXISTS `ref`;
CREATE TABLE `ref` (
  `Id` int(11) NOT NULL AUTO_INCREMENT,
  `key1` int(11) DEFAULT '-99',
  `key2` int(11) DEFAULT '-99',
  PRIMARY KEY (`Id`),
  KEY `main` (`key1`,`key2`)
) ENGINE=MyISAM AUTO_INCREMENT=7 DEFAULT CHARSET=latin1;


LOCK TABLES `ref` WRITE;
INSERT INTO `ref` VALUES (1,1,1),(2,1,2),(3,2,2),(4,3,1),(5,3,2),(6,3,3);
UNLOCK TABLES;

【问题讨论】：

标签： mysql sql join query-optimization mariadb

【解决方案1】：

您可以从dts 表中取消透视数据，然后在其上使用group_concat。

SELECT 
    r.key1,
    r.key2,
    group_concat(distinct case when val > 0 then pr end order by pr separator ',') prs
FROM
    ref r
        INNER JOIN
    (
    SELECT
    d.key1,
    d.key2,
    t.pr,
    CASE t.pr
        WHEN 'pr1' THEN pr1
        WHEN 'pr2' THEN pr2
        WHEN 'pr3' THEN pr3
        WHEN 'pr4' THEN pr4
        WHEN 'pr5' THEN pr5
    END val
FROM
    dts d
        CROSS JOIN
    (
    SELECT 'pr1' pr UNION ALL 
    SELECT 'pr2' UNION ALL 
    SELECT 'pr3' UNION ALL 
    SELECT 'pr4' UNION ALL 
    SELECT 'pr5') t
    ) d ON r.key1 = d.key1 AND r.key2 = d.key2
GROUP BY r.key1 , r.key2;

生产：

+------+------+---------------------+
| key1 | key2 | prs                 |
+------+------+---------------------+
|    1 |    1 | pr1,pr3,pr4,pr5     |
|    1 |    2 | pr2,pr4,pr5         |
|    2 |    2 | pr1,pr2,pr3,pr4,pr5 |
|    3 |    2 | NULL                |
|    3 |    3 | pr1,pr2,pr3,pr4,pr5 |
+------+------+---------------------+
5 rows in set (0.00 sec)

编辑：

没有加入ref 表（因为 ref 表有所有的 key1、key2 而我们只是在内部加入）：

SELECT 
    key1,
    key2,
    group_concat(distinct case when val > 0 then pr end order by pr separator ',') prs
FROM (
    SELECT 
        d.key1,
            d.key2,
            t.pr,
            CASE t.pr
                WHEN 'pr1' THEN pr1
                WHEN 'pr2' THEN pr2
                WHEN 'pr3' THEN pr3
                WHEN 'pr4' THEN pr4
                WHEN 'pr5' THEN pr5
            END val
    FROM
        dts d
    CROSS JOIN (
        SELECT 'pr1' pr UNION ALL 
        SELECT 'pr2' UNION ALL 
        SELECT 'pr3' UNION ALL 
        SELECT 'pr4' UNION ALL 
        SELECT 'pr5'
    ) t
) r
GROUP BY key1 , key2;

产生相同的输出：

+------+------+---------------------+
| key1 | key2 | prs                 |
+------+------+---------------------+
|    1 |    1 | pr1,pr3,pr4,pr5     |
|    1 |    2 | pr2,pr4,pr5         |
|    2 |    2 | pr1,pr2,pr3,pr4,pr5 |
|    3 |    2 | NULL                |
|    3 |    3 | pr1,pr2,pr3,pr4,pr5 |
+------+------+---------------------+
5 rows in set (0.00 sec)

编辑 2：

SELECT 
    r.key1,
    r.key2,
    group_concat(distinct case when val > 0 then pr end order by pr separator ',') prs
FROM (
    select key1, key2
    from ref
    order by id
    limit 0, 1000       -- Added limit to get only first 1000 key pairs based on id
) r INNER JOIN (
    SELECT
    d.key1,
    d.key2,
    t.pr,
    CASE t.pr
        WHEN 'pr1' THEN pr1
        WHEN 'pr2' THEN pr2
        WHEN 'pr3' THEN pr3
        WHEN 'pr4' THEN pr4
        WHEN 'pr5' THEN pr5
    END val
FROM
    dts d
        CROSS JOIN
    (
    SELECT 'pr1' pr UNION ALL 
    SELECT 'pr2' UNION ALL 
    SELECT 'pr3' UNION ALL 
    SELECT 'pr4' UNION ALL 
    SELECT 'pr5') t
    ) d ON r.key1 = d.key1 AND r.key2 = d.key2
GROUP BY r.key1 , r.key2;

对于前 1000 个唯一密钥对，在上述查询中使用以下 SQL：

(
    select key1, key2
    from ref
    group by key1, key2
    order by key1, key2
    limit 0, 1000       -- Added limit to get only first 1000 key pairs based on id
) r

【讨论】：

它给出了` | 3 | 1 | NULL |, key1=3` 和 key2=1 在 dts 表中不存在，但仍显示在结果中
^1 是的我试过它的工作，试图理解不能进一步简化？
没有参考表怎么办？
为什么是SELECT 'pr1' pr UNION ALL？
这是取消透视表的技巧 - 即将列转换为行。与 oracle 或 sql-server 等其他 DBMS 不同，MySQL 没有内置支持。