Django多个注释会减慢查询速度 [英] Django multiple annotate slows down the query

问题描述

我一直在使用 django debug_toolbar 进行调试，如果我在查询中使用多个注释，那么Django需要花费很多时间来获取查询结果。

I have been debugging with django debug_toolbar, if I use more than one annotate in query then it takes a lot of time for Django to fetch the query results.

class Project_First(models.Model):
   project_first_results_M2M = models.ManyToManyField(Project_First_Results)

class Project_Second(models.Model):
   project_second_results_M2M = models.ManyToManyField(Project_Second_Results)

class Project(models.Model):
    project_first_M2M = models.ManyToManyField(Project_First)
    project_second_M2M = models.ManyToManyField(Project_Second)

• 我正在尝试计算所有 project_first_M2M 对象中 project_first_results_M2M 中存在的所有对象。


• I m trying to count all the objects present in project_first_results_M2M of all the project_first_M2M objects.

即，让我们假设 project_first_M2M 具有3个<$ c $的对象c> Project_First ，我想计算所有这三个对象中所有 project_first_results_M2M 个对象。

ie, let's suppose, project_first_M2M has 3 objects of Project_First and I want to count all the total project_first_results_M2M objects present in all 3 of them.

Project.objects.all().annotate(first_res_count=Count('project_first_M2M__project_first_results_M2M',distinct=True))







• 上面的查询工作正常，需要80毫秒来获取结果。但是当我尝试添加其他注释进行查询时，就会出现问题。


• Above Query works fine and would take 80ms to fetch the results. But problem occurs when i try to add an additional annotate to query.

Project.objects.all().annotate(first_res_count=Count('project_first_M2M__project_first_results_M2M',distinct=True)).annotate(second_res_count=Count('project_second_M2M__project_second_results_M2M',distinct=True))







• 这将花费近4000毫秒来获取结果。



project_second_M2M 和 project_first_M2M 都包含相同的内容字段和相同数量的对象。反之，我甚至尝试了上述查询，并且只有当我添加其他注释时，查询速度才会降低。

Both project_second_M2M and project_first_M2M contains the same fields and the same number of objects. I even tried in vice-versa condition to the above query and query slows down only when I add additional annotate.



• 是否有任何快速替代的解决方案以高效的方式实现相同目标？


• 我想计算所有 project_first_M2M中 project_first_results_M2M 的所有对象每个 Project 对象中的code>个对象，对于 project_second_results_M2M

ul>

• Is there any fast and alternate solution to achieve the same in much efficient way? maybe with raw sql queries.
• I want to count all the objects of project_first_results_M2M of all project_first_M2M objects within each Project object and similarly for project_second_results_M2M

推荐答案

可能您可以使用 与预取有关 ：

Project.objects.prefetch_related('project_first_M2M__project_first_results_M2M', 'project_second_M2M__project_second_results_M2M').annotate(first_res_count=Count('project_first_M2M__project_first_results_M2M',distinct=True)).annotate(second_res_count=Count('project_second_M2M__project_second_results_M2M',distinct=True))

这篇关于Django多个注释会减慢查询速度的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！