汇总数字，一些在范围内

Question

我有一个表tab1，在我的数字列c1 中有以下值：

1
2
3
4
5
6
7
8
9
10
13
18

我想回来

1-5
6-10
13
18

下面的查询在 MySQL 而不是 PostgreSQL 中成功执行：

select distinct
case
when c1 between 1 and 5 then '1-5'
when c1 between 6 and 9 then '6-10'
else c1
end as 'c1'
from tab1;

Answer 1

修复语法错误并正确转换后，相同的查询在 Postgres 中有效：

SELECT DISTINCT
       CASE WHEN c1 BETWEEN 1 AND 5 THEN '1-5'
            WHEN c1 BETWEEN 6 AND 9 THEN '6-10'
            ELSE c1::text END AS c1 
FROM   tab1;

结果必须是字符串类型 - text 是 Postgres 中的首选类型。但是按字符串表示排序对您的数字没有好处。所以我建议这种替代方法允许正确排序：

SELECT CASE WHEN c1 BETWEEN 1 AND 5 THEN '1-5'
            WHEN c1 BETWEEN 6 AND 9 THEN '6-10'
            ELSE c1::text END AS new_c1   -- avoid duplicate of source name
FROM   tab1 t
GROUP  BY new_c1
ORDER  BY min(t.c1);

db小提琴here

Answer 2

如果我理解正确，您将使用两个级别的聚合。第一个获取范围：

select min(c1), max(c1)
from (select t.*, row_number() over (order by c1) as seqnum
      from t
     ) t
group by c1 - seqnum;

第二个然后重新聚合：

select string_agg(range, ',' order by min_c1)
from (select (case when min(c1) <> max(c1) then min(c1) || '-' || max(c1) else min(c1)::text end) as range, min(c1) as min_c1
      from (select t.*, row_number() over (order by c1) as seqnum
            from t
           ) t
      group by c1 - seqnum
     ) x

Here 是一个 dbfiddle，说明此逻辑有效。