Question

我有这个查询，我正在尝试使用Django ORM重新创建。

select
    count(tickets.id),
    date
from
    tickets,
    generate_series('2000-01-01', date_trunc('month', CURRENT_DATE), interval '1 month') as date
where
    (date >= created AND date < resolved)
group by
    date;

我想我很接近这个：

tkts = Ticket.objects.extra(
    tables=["generate_series('2000-01-01', date_trunc('month', CURRENT_DATE), interval '1 month') as date"],
    where=["(date >= created AND date < resolved)"],
    order_by=['date'],
).values('date').annotate(cnt=Count('id'))

但我收到错误：

FieldError：无法将关键字'date'解析为字段。

当我将查询减少到此时：

tkts = Ticket.objects.extra(
    tables=["generate_series('2000-01-01', date_trunc('month', CURRENT_DATE), interval '1 month') as date"],
    where=["(date >= created AND date < resolved)"])

我可以看到表中的值被转义并用双引号括起来（来自connections[app].queries：

'SELECT ...
FROM 
  "tickets",
  "generate_series(\'2000-01-01\', date_trunc(\'month\', CURRENT_DATE), interval \'1 month\') as date"
WHERE
 ("tickets"."queue" IN (1,2,3) AND
 ((date >= created AND date < resolved)))
ORDER BY "tickets"."created" DESC LIMIT 21'

那么我可以用mark_safe来避免这种情况吗？我想我可以创建一个存储过程来替换generate_series的东西，但如果可能的话我宁愿不改变数据库。

使用Django ORM聚合生成的系列

0 个答案: