annotate分组

# 原始SQL
SELECT start_date, plat_account, site_code, SUM(sessions) AS sessions  FROM crl_reports_business_reports WHERE start_date >= '2021-9-30'  GROUP BY plat_account, start_date, site_code  ORDER BY plat_account, site_code, start_date;

# ORM
CrlReportsBusinessReports.objects.filter(start_date__gte=datetime.datetime.now().date() - datetime.timedelta(days=20)).values('plat_account', 'site_code', 'start_date').annotate(count_=Sum('sessions')).order_by('plat_account', 'site_code', 'start_date')

# ORM生成的SQL
SELECT `crl_reports_business_reports`.`plat_account`, `crl_reports_business_reports`.`site_code`, `crl_reports_business_reports`.`start_date`, SUM(`crl_reports_business_reports`.`sessions`) AS `count_` FROM `crl_reports_business_reports` WHERE `crl_reports_business_reports`.`start_date` >= 2021-09-29 GROUP BY `crl_reports_business_reports`.`plat_account`, `crl_reports_business_reports`.`site_code`, `crl_reports_business_reports`.`start_date` ORDER BY `crl_reports_business_reports`.`plat_account` ASC, `crl_reports_business_reports`.`site_code` ASC, `crl_reports_business_reports`.`start_date` ASC

aggregate聚合

很简单就是查询一个数据集的某个字段的最大值、最小值、平均数、求和、计数等，返回值是字典格式