在学习Django聚合与分组查询中，发现value与annotate的顺序不同时，查询结果大相径庭，经过一下午的研究，终于弄明白了，现在分享给大家，先上结论：
结论
- value在annotate前面时，相当于
group by，即models.表名.objects.values(\'group by 字段\').annotate(统计字段) - value放在annotate后面时，相当于
select - annotate前面没有value时，默认按当前表的主键分组
验证
表格准备
查询
value放在前面时
ret=models.Student.objects.values(\'gender\').annotate(count=Count(\'sid\')).order_by(\'-count\')
print(ret)
等价于以下sql语句
SELECT `app_student`.`gender`, COUNT(`app_student`.`sid`) AS `count` FROM `app_student` GROUP BY `app_student`.`gender` ORDER BY `count` DESC
执行结果
<QuerySet [{\'gender\': \'女\', \'count\': 6}, {\'gender\': \'男\', \'count\': 5}]>
可以看到结果是按性别分组的
### value放在后面时
ret = models.Student.objects.annotate(count=Count(\'sid\')).values(\'gender\').order_by(\'-count\')
print(ret)
等价于以下sql语句
SELECT `app_student`.`gender` FROM `app_student` GROUP BY `app_student`.`sid` ORDER BY COUNT(`app_student`.`sid`) DESC
执行结果
<QuerySet [{\'gender\': \'女\'}, {\'gender\': \'男\'}, {\'gender\': \'女\'}, {\'gender\': \'女\'}, {\'gender\': \'男\'}, {\'gender\': \'女\'}, {\'gender\': \'男\'}, {\'gender\': \'男\'}, {\'gender\': \'女\'}, {\'gender\': \'女\'}, {\'gender\': \'男\'}]>
可以看到结果并没有按性别分组，而是每个学生作为一组，说明上述结论是正确的