Question

我有以下查询，执行时间很长（大约30秒）。有没有办法优化这个查询或者我应该让它成为？

{{1}}

Answer 1

您应该在用于连接的所有列上都有索引。更重要的是，您应该在db.vehicle上有一个索引：

db_vehicle(name, model, engine_type, id)

这应该用于where子句并加快查询速度。

Answer 2

正如@Gordon Linoff已经指出的那样，你需要定义索引。但是，您的查询还有一些事情要做。定义好表格的顺序当然更好，所以我们改变了这个

SELECT l.Name, l.Model, l.Engine_Type, m.ID, g.Name  
FROM db.vehicle_part k
JOIN db.vehicle l ON k.ID_Vehicle = l.ID
JOIN db.part m ON k.ID_Part = m.ID

JOIN db.part_generic_part b ON b.ID_Part = m.ID
JOIN db.generic_part c ON c.ID = b.ID_Generic_part
JOIN db.generic_part_name d ON c.ID = d.ID_Generic_Part 
JOIN db.part_category_generic_part e ON c.ID = e.ID_generic_part
JOIN db.part_category f ON e.ID_category = f.ID
JOIN db.part_category_name g ON f.ID = g.ID_Part_Category 
JOIN db.lang h ON g.ID_Lang = h.ID

WHERE l.Name = 'BMW' AND l.Model = '3 (E30)' AND l.Engine_Type= '316 i' AND g.ID_Lang =2

到这个

SELECT l.Name, l.Model, l.Engine_Type, m.ID, g.Name  
FROM db.vehicle_part k
JOIN db.vehicle l ON k.ID_Vehicle = l.ID
JOIN db.part m ON k.ID_Part = m.ID

JOIN db.part_generic_part b ON b.ID_Part = m.ID
JOIN db.generic_part c ON c.ID = b.ID_Generic_part
JOIN db.part_category_generic_part e ON c.ID = e.ID_generic_part
JOIN db.part_category f ON e.ID_category = f.ID
JOIN db.part_category_name g ON f.ID = g.ID_Part_Category 
JOIN db.lang h ON g.ID_Lang = h.ID
JOIN db.generic_part_name d ON c.ID = d.ID_Generic_Part 

WHERE l.Name = 'BMW' AND l.Model = '3 (E30)' AND l.Engine_Type= '316 i' AND g.ID_Lang =2

这是一个更优雅的代码，即使不能保证它实际上比初始代码更快。但是，通过这个更好的订单，我们进入下一步。 where子句在join完成后执行，因此您将大量多维记录加载到内存中并过滤掉行。 on条件之前执行，因此下一步是修改where并将其条件迁移到on条件，如下所示：

SELECT l.Name, l.Model, l.Engine_Type, m.ID, g.Name  
FROM db.vehicle_part k
JOIN db.vehicle l ON k.ID_Vehicle = l.ID AND l.Name = 'BMW' AND l.Model = '3 (E30)' AND l.Engine_Type= '316 i'
JOIN db.part m ON k.ID_Part = m.ID

JOIN db.part_generic_part b ON b.ID_Part = m.ID
JOIN db.generic_part c ON c.ID = b.ID_Generic_part
JOIN db.part_category_generic_part e ON c.ID = e.ID_generic_part
JOIN db.part_category f ON e.ID_category = f.ID
JOIN db.part_category_name g ON f.ID = g.ID_Part_Category and g.ID_Lang =2
JOIN db.lang h ON g.ID_Lang = h.ID
JOIN db.generic_part_name d ON c.ID = d.ID_Generic_Part

这样，在join时过滤掉了不需要的记录，减少了时间，因为它们并非都被加载到内存中然后被过滤掉了。当且仅当查询仍然很慢时，应该完成最后一步，因为它降低了代码的可读性：您可以修改具有严格过滤器的join（大幅减少行数）或严重投影（具有许多大字段，但仅需要其中的一小部分），以命名子选择。例如：

SELECT 'BMW', '3 (E30)', '316 i', m.ID, g.Name  
from (select temp.ID from db.vehicle temp where temp.Name = 'BMW' AND temp.Model = '3 (E30)' AND temp.Engine_Type= '316 i') l
JOIN db.vehicle_part k on l.ID = k.ID_VEHICLE
JOIN db.part m ON k.ID_Part = m.ID

JOIN db.part_generic_part b ON b.ID_Part = m.ID
JOIN db.generic_part c ON c.ID = b.ID_Generic_part
JOIN db.part_category_generic_part e ON c.ID = e.ID_generic_part
JOIN db.part_category f ON e.ID_category = f.ID
JOIN db.part_category_name g ON f.ID = g.ID_Part_Category and g.ID_Lang =2
JOIN db.lang h ON g.ID_Lang = h.ID
JOIN db.generic_part_name d ON c.ID = d.ID_Generic_Part

此最终修改预过滤l以确保只有少数记录可以启动join。

需要MySQL查询优化帮助

2 个答案: