Question

强文本，我遇到以下问题：给定两个表的联系人和组织：

WITH contacts(oe_id, name, email, person_id) AS (VALUES
    (1, 'Mark', 'm.smith@test.nl', 19650728),
    (2, 'Tom', 't.b.smith@test.nl', 20010627), 
    (1, 'Frank', 'f.j.smith@test.nl', 20040709), 
    (3, 'Petra', 'p.ringenaldus@test.nl', 19700317), 
    (3, 'Paul', 'p.m.sprengers@test.nl', 19681006)),
organisations(oe_id, name) AS (VALUES
    (1, 'Cardiology'), 
    (2, 'Neurology'), 
    (3, 'Dermatology'), 
    (4, 'Churgery'))

我想得到一个包含3列的表格：组织名称，组织ID和该组织的联系人数组。每个数组元素也是一个包含联系人数据的数组。

首先，我创建了一个表，其中所有联系人列都被汇总到一个数组中。每元组行一个数组：

 WITH contacts(oe_id, name, email, person_id) AS (VALUES
    (1, 'Mark', 'm.smith@test.nl', 19650728), 
    (2, 'Tom', 't.b.smith@test.nl', 20010627), 
    (1, 'Frank', 'f.j.smith@test.nl', 20040709), 
    (3, 'Petra', 'p.ringenaldus@test.nl', 19700317), 
    (3, 'Paul', 'p.m.sprengers@test.nl', 19681006)),
organisations(oe_id, name) AS (VALUES
    (1, 'Cardiology'), 
    (2, 'Neurology'), 
    (3, 'Dermatology'), 
    (4, 'Churgery')),
contacts_aggregated(oe_id, cdata) AS (
    select oe_id, ARRAY[name, email, person_id::text] from contacts) 
select * from contacts_aggregated;

此结果为：

 oe_id |                 cdata
-------+---------------------------------------
     1 | {Mark,m.smith@test.nl,19650728}
     2 | {Tom,t.b.smith@test.nl,20010627}
     1 | {Frank,f.j.smith@test.nl,20040709}
     3 | {Petra,p.ringenaldus@test.nl,19700317}
     3 | {Paul,p.m.sprengers@test.nl,19681006}
(5 rows)

下一步是为每个组织ID汇总cdata（联系数据）：

WITH contacts(oe_id, name, email, person_id) AS (VALUES
    (1, 'Mark', 'm.smith@test.nl', 19650728), 
    (2, 'Tom', 't.b.smith@test.nl', 20010627), 
    (1, 'Frank', 'f.j.smith@test.nl', 20040709), 
    (3, 'Petra', 'p.ringenaldus@test.nl', 19700317), 
    (3, 'Paul', 'p.m.sprengers@test.nl', 19681006)),
organisations(oe_id, name) AS (VALUES
    (1, 'Cardiology'), 
    (2, 'Neurology'), 
    (3, 'Dermatology'), 
    (4, 'Churgery')),
contacts_aggregated(oe_id, cdata) AS (
    select oe_id, ARRAY[name, email, person_id::text] from contacts),
contacts_for_organisations(oe_id, contacts) AS (
    SELECT organisations.oe_id, array_agg(contacts_aggregated.cdata::text) 
    FROM organisations
    JOIN contacts_aggregated USING(oe_id) 
    GROUP BY oe_id)
SELECT * FROM contacts_for_organisations;

结果如下：

 oe_id |                                      contacts
-------+------------------------------------------------------------------------------------
     1 | {"{Mark,m.smith@test.nl,19650728}","{Frank,f.j.smith@test.nl,20040709}"}
     2 | {"{Tom,t.b.smith@test.nl,20010627}"}
     3 | {"{Petra,p.ringenaldus@test.nl,19700317}","{Paul,p.m.sprengers@test.nl,19681006}"}
(3 rows)

如您所见，结果是一个数组。但是它的元素也应该是一个数组。代替数组，我得到一个内爆数组作为字符串。

我想要的Wat是这样的：

 oe_id |                                      contacts
-------+------------------------------------------------------------------------------------
     1 | {{Mark,m.smith@test.nl,19650728},{Frank,f.j.smith@test.nl,20040709}}
     2 | {{Tom,t.b.smith@test.nl,20010627}}
     3 | {{Petra,p.ringenaldus@test.nl,19700317},{Paul,p.m.sprengers@test.nl,19681006}}
(3 rows)

如果我删除对文本array_agg（contacts_aggregated.cdata :: text的转换，我会得到：

could not find array type for data type text[]

我忘记/做错了什么？ Postgres：psql（9.2.24）和psql（9.6.10，服务器9.2.24）

如果我在postgres 9.6服务器上使用postgres客户端9.6运行代码，则一切正常。

Answer 1

我刚刚移至更高的postgres版本，现在一切正常。

Postgres查询应返回一个数组而不是字符串数组

1 个答案: