Postgres结果多维查询

Question

How would one create a Multidimensional Array from a single Postgres Query.

I have 3 tables with the following columns:

• tb_school (id, school_name) eg:

  [ {id:"1", school_name:"School1"}, {id:"2", school_name:"School2"} ]

• tb_profile (id, profile_name, school_id) eg:

  [ {id:"1", profile_name:"John", school_id:"1"}, {id:"2", profile_name:"Peter", school_id:"1"}, {id:"3", profile_name:"Sam", school_id:"1"}, {id:"4", profile_name:"Susan", school_id:"2"}, {id:"5", profile_name:"Jude", school_id:"2"}, {id:"6", profile_name:"Kim", school_id:"2"} ]

• tb_article (id, article_name, profile_id) eg:

  [ {id:"1", article_name:"Headline News", profile_id:"1"}, {id:"2", article_name:"Sports Recap", profile_id:"2"}, {id:"3", article_name:"Weather", profile_id:"3"}, {id:"4", article_name:"Arts", profile_id:"4"}, {id:"5", article_name:"Other", profile_id:"5"}, {id:"6", article_name:"Example", profile_id:"6"} ]

I would like the query to return a result lat looks something like this:

[
    {school_name:"School1", people:[
        {profile_name:"John", articles:[
            article_name:"Headline News"
        ]},
        {profile_name:"Peter", articles:[
            article_name:"Sports Recap"
        ]},
        {profile_name:"Sam", articles:[
            article_name:"Weather"
        ]},
    ]},
    {school_name:"School2", people:[
        {profile_name:"Susan", articles:[
            article_name:"Arts"
        ]},
        {profile_name:"Jude", articles:[
            article_name:"Other"
        ]},
        {profile_name:"Kim", articles:[
            article_name:"Example"
        ]},
    ]}
]

I know one can achieve this by 3 separate select queries in nested for loops. What I want to know is could this be achieved by one single postgres query?

Answer 1

Note: I am not good with PostgreSQL. Probably made wrong assumptions and code is probably not intended for production.

I almost sure you cannot just result a nested object. But you can receive JSON. So

 SELECT json_build_object( 'school_name', school_name, 'people', ( SELECT json_agg(json_build_object( 'profile_name', profile_name, 'articles', ( SELECT json_agg(json_build_object( 'article_name', article_name )) FROM tb_article WHERE profile_id = tb_profile.facid ) )) FROM tb_profile WHERE school_id = tb_school.id ) ) AS result FROM tb_school 

it will give you row as JSON, which you can JSON.parse .

If you want to test you can use this learning service . I used it with this code.

 SELECT json_build_object( 'name', firstname, 'slots', ( SELECT json_agg(json_build_object( 'number', slots, 'facility', ( SELECT json_agg(json_build_object( 'name', name )) FROM cd.facilities WHERE facid = cd.bookings.facid ) )) FROM cd.bookings WHERE memid = cd.members.memid ) ) AS result FROM cd.members 

It probably will be helpful or provide more ideas.