1

我需要从 tableA 中查询前 5 个值。比如下面

    select id, count(occurrence), date 
    from 
    (select id, unnest(value) as occurrence, date from tableA) as a
    group by id, occurrence, date 
    order by occurrence desc 
    limit 5

 id  | occurrence |   date   
-----+-------+----------
 330 |    11 | 20141015
 400 |    11 | 20141015
 390 |    10 | 20141015
 240 |    10 | 20141015
 501 |    10 | 20141015

并在收到 id 后,查询同一个 tableA 以获取其他日期范围从 20140101 到 20141015 的 id 的所有值。

expected result:

 id  | occurrence |   date   
-----+-------+----------
 330 |    11 | 20141015
 400 |    11 | 20141015
 390 |    10 | 20141015
 240 |    10 | 20141015
 501 |    10 | 20141015
 330 |    0  | 20141014
 400 |    1  | 20141014
 390 |    10 | 20141014
 240 |    15 | 20141014
 501 |    10 | 20141014
 330 |    11 | 20141013
 400 |    11 | 20141013
 390 |    11 | 20141013
 240 |    19 | 20141013
 501 |    10 | 20141013

但我究竟该怎么做呢?

我的 postgresql 版本是 8.1,我不能使用分区(如果你想建议的话)


编辑

select  id, count(value) as occurrence, date
from tableA
where id = ANY(
    select array(
        select id
        from (
            select date, unnest(id) as id
            from tableA where date>='20140101' and date<='20141015'
            )as a
        )
)
group by id, date

不返回任何东西。我的数组正确吗?

4

2 回答 2

0

您可以尝试 select sth from table where dt > 20140101 and dt < 20141015 and row_num <= 5 这会给你正确的答案。

于 2015-10-13T14:51:48.567 回答
0
select id, count(value) as occurrence, date
from t
where id in (
    select id
    from (
        select id, count(value) as occurrence, date 
        from t 
        group by id, date 
        order by occurrence desc 
        limit 5
    ) s
) q
group by id, date
于 2015-10-13T14:53:37.650 回答