[TimLinux] django aggregate和annotate示例

1. 聚合与注解

聚合(aggregate)比较好理解，注解(annotate)真不好理解，这篇示例参考了文章“django中聚合aggregate和annotate GROUP BY的使用方法”提供的模型，以及部分内容。根据参考的文章，注解的理解是：在聚合的基础上使用了GROUP BY语句。

2. 模型代码

from django.db import models

class Author(models.Model):
    name = models.CharField(max_length=100)
    age = models.IntegerField()

    class Meta:
        db_table = 'author'

class Publisher(models.Model):
    name = models.CharField(max_length=300)
    num_awards = models.IntegerField()

    class Meta:
        db_table = 'publisher'

class Book(models.Model):
    name = models.CharField(max_length=300)
    pages = models.IntegerField()
    price = models.DecimalField(max_digits=10, decimal_places=2)
    rating = models.FloatField()
    authors = models.ManyToManyField(Author)
    publisher = models.ForeignKey(Publisher, on_delete=models.CASCADE)
    pubdate = models.DateField()

    class Meta:
        db_table = 'book'

class Store(models.Model):
    name = models.CharField(max_length=300)
    books = models.ManyToManyField(Book)
    registered_users = models.PositiveIntegerField()

    class Meta:
        db_table = 'store'

3. aggregate

# 需要导入的模块
>>> from django.db.models import Count, Avg, Sum, F, FloatField
>>> from django.db import connections
>>> from books.models import *

#1. 聚合返回的是字典值
统计计数
>>> Publisher.objects.aggregate(num_books=Count('book'))
>>> connections['default'].queries
SELECT
    COUNT(`book`.`id`) AS `num_books`
FROM
    `publisher` LEFT OUTER JOIN `book`
    ON (`publisher`.`id` = `book`.`publisher_id`);

求平均
>>> Book.objects.all().aggregate(Avg('price'))
>>> Book.objects.aggregate(Avg('price'))
SELECT
    AVG(`book`.`price`) AS `price__avg`
FROM
    `book`;

求每页价格
>>> Book.objects.aggregate( price_per_page=Sum(F('price')/F('pages'), output_field=FloatField()) )
SELECT
    SUM((`book`.`price` / `book`.`pages)) AS `price_per_page`
FROM
　　`book`; 

4. annotate

对数据进行分组后，再聚合，使用annotate

# 导入涉及的模块
from django.db.models import Count, Sum, FloatField
from django.db import connections
from books.models import *

# annotate是分组、聚合的含义，annotate返回的是新的queryset
# aggregate则是对全表进行聚合的含义，没有分组的概念，返回的是字典
>>> Book.objects.annotate(Count('authors'))  # authors是Book中的属性
SELECT
    `book`.`id`, `book`.`name`, `book`.`pages`, `book`.`price`,
    `book`.`rating`, `book`.`publisher_id`, `book`.`pubdate`,
    COUNT(`book_authors`.`author_id`) AS `authors__count`
FROM
    `book`
LEFT OUTER JOIN
    `book_authors`
ON
    (`book`.`id` = `book_authors`.`book_id`)
GROUP BY
    `book`.`id`
ORDER BY
    NULL;