SQL加入大数据

Question

我有这个加入：

SELECT a.* 
FROM `entities_instance` a  
 INNER JOIN `instance_fields_data` artists 
 ON ( a.entity_instance_id = artists.entity_instance_id 
      AND artists.`dynamic_field_name` = 'artists' 
      AND artists.`field_value` like '%8605:%')
 INNER JOIN `instance_fields_data` _enddate 
 ON (a.entity_instance_id = _enddate.entity_instance_id) 
    AND (_enddate.`dynamic_field_name` = 'enddate') 
WHERE a.`entity_id` = 67 
GROUP BY a.entity_instance_id
ORDER BY _enddate.field_value DESC 
LIMIT 100

它需要返回大约5个结果，但我得到了这个错误：

The SELECT would examine more than MAX_JOIN_SIZE rows; 
check your WHERE and use SET SQL_BIG_SELECTS=1 or 
SET MAX_JOIN_SIZE=# if the SELECT is okay

有人可以帮我改进这个查询吗？ 我不想使用SQL_BIG_SELECTS命令。

谢谢

Answer 1

网上的许多资源建议在主查询之前将SQL_BIG_SELECTS=1作为单独的查询运行。

我在想，也许您可​​以尝试使用别名表来执行较小数据大小的连接。不确定，但它可以帮助。

示例：

select a.* from
(select * from `entities_instance` test WHERE test.`entity_id` = 67) as a
INNER JOIN `instance_fields_data` artists 
 ON ( a.entity_instance_id = artists.entity_instance_id 
      AND artists.`dynamic_field_name` = 'artists' 
      AND artists.`field_value` like '%8605:%')
 INNER JOIN `instance_fields_data` _enddate 
 ON (a.entity_instance_id = _enddate.entity_instance_id) 
    AND (_enddate.`dynamic_field_name` = 'enddate') 
GROUP BY a.entity_instance_id
ORDER BY _enddate.field_value DESC 
LIMIT 100;

为了提供更多帮助，您可以获得数据和表格结构的示例。 例如。我们甚至不知道哪个表是数据最多的表。

<强>例2：

这里我将为3个表中的每个表获取where语句，并在别名派生表中使用它们。

  SELECT a.*
FROM
  (SELECT *
   FROM `entities_instance` test
   WHERE test.`entity_id` = 67) AS a,

  (SELECT *
   FROM `instance_fields_data` test2
   WHERE test2.`dynamic_field_name` = 'artists'
     AND test2.`field_value` LIKE '%8605:%') AS artists,

  (SELECT *
   FROM `instance_fields_data` test3
   WHERE test3.`dynamic_field_name` = 'enddate') AS _enddate
WHERE a.entity_instance_id = artists.entity_instance_id
  AND a.entity_instance_id = _enddate.entity_instance_id
GROUP BY a.entity_instance_id
ORDER BY _enddate.field_value DESC LIMIT 100;

<强>示例3：

最后，在此进一步优化中，我们使用EXISTS来进一步过滤instance_fields_data的两个子查询的结果。我确信这可以进一步简化/优化。还要确保您的密钥上有索引（例如entity_instance_id），否则这将更慢。

SELECT a.*
FROM
  (SELECT *
   FROM `entities_instance` test
   WHERE test.`entity_id` = 67) AS a,

  (SELECT *
   FROM `instance_fields_data` test2
   WHERE test2.`dynamic_field_name` = 'artists'
     AND test2.`field_value` LIKE '%8605:%'
     AND EXISTS
       (SELECT *
        FROM `entities_instance` z
        WHERE z.entity_instance_id = test2.entity_instance_id AND
        z.`entity_id` = 67)) AS artists,

  (SELECT *
   FROM `instance_fields_data` test3
   WHERE test3.`dynamic_field_name` = 'enddate'
     AND EXISTS
       (SELECT *
        FROM `entities_instance` z
        WHERE z.entity_instance_id = test3.entity_instance_id AND
        z.`entity_id` = 67)) AS _enddate
WHERE a.entity_instance_id = artists.entity_instance_id
  AND a.entity_instance_id = _enddate.entity_instance_id
GROUP BY a.entity_instance_id
ORDER BY _enddate.field_value DESC LIMIT 100;