Question

数据库是PostgreSQL 9.6。

create table candidates (
  id            int             not null,
  name          varchar,
  email         varchar,
  created_at    timestamptz,
  updated_at    timestamptz
);

insert into candidates values
  (1,   'John Snow',        'john@snow',    now()+interval '1 seconds', now()+interval '1 seconds'),
  (2,   'Bob Ross',         'bob@ross',     now()+interval '2 seconds', now()+interval '2 seconds'),
  (3,   'Kim Possible',     'kim@possible', now()+interval '3 seconds', now()+interval '3 seconds'),
  (4,   'Ash Ketchum',      'ash@ketchum',  now()+interval '4 seconds', now()+interval '4 seconds'),
  (5,   'Goku Son',         'goku@son',     now()+interval '5 seconds', now()+interval '5 seconds'),
  (6,   'Mortimer Smith',   'morty@smh',    now()+interval '6 seconds', now()+interval '6 seconds');

create table jobs (
  id            int             not null,
  title         varchar,
  positions     int,
  created_at    timestamptz,
  updated_at    timestamptz
);

insert into jobs values
  (1,   'Night Watch Commander',    1,      now()+interval '11 seconds',    now()+interval '11 seconds'),
  (2,   'Pokemon Master',           8,      now()+interval '12 seconds',    now()+interval '12 seconds'),
  (3,   'Blessed Painter',          100,    now()+interval '13 seconds',    now()+interval '13 seconds');

create table applications (
  id            int             not null,
  candidate_id  int             not null,
  job_id        int             not null,
  score         decimal,
  created_at    timestamptz,
  updated_at    timestamptz
);

insert into applications values
  (1,   1,  1,  90,     now()+interval '21 seconds',    now()+interval '21 seconds'),
  (2,   1,  3,  80,     now()+interval '22 seconds',    now()+interval '22 seconds'),
  (3,   2,  1,  10,     now()+interval '23 seconds',    now()+interval '23 seconds'),
  (4,   2,  2,  10,     now()+interval '24 seconds',    now()+interval '24 seconds'),
  (5,   2,  3,  99,     now()+interval '25 seconds',    now()+interval '25 seconds'),
  (6,   3,  1,  60,     now()+interval '26 seconds',    now()+interval '26 seconds'),
  (7,   3,  2,  70,     now()+interval '27 seconds',    now()+interval '27 seconds'),
  (8,   3,  3,  50,     now()+interval '28 seconds',    now()+interval '28 seconds'),
  (9,   4,  2,  90,     now()+interval '29 seconds',    now()+interval '29 seconds'),
  (10,  5,  1,  9000,   now()+interval '30 seconds',    now()+interval '30 seconds');

对于至少具有一个应用程序的每个候选人，查询要求提供姓名，其最大应用程序分数的职位名称以及该应用程序的分数。（如果有关系，请选择最后创建的应用程序）。要进行测试，请使用网站http://sqlfiddle.com

Expected output:
name             title                   score
Ash Ketchum      Pokemon Master          90
Bob Ross         Blessed Painter         99
Goku Son         Night Watch Commander   9000
John Snow        Night Watch Commander   90
Kim Possible     Pokemon Master          70

Answer 1

按分数排序的第一位，按候选人划分。这就是我在下面的“ rankApps”中所做的。然后为每个候选人选择排名最高的职位，加入候选人和职位，然后输出。

with 

    rankApps as (

        select     candidate_id, job_id, score,
                   row_number() over(partition by candidate_id order by score desc) as rnk
        from       applications

    )

    select        c.name, j.title, a.score
    from          rankApps a
    join          candidates c on a.candidate_id = c.id 
    join          jobs j on a.job_id = j.id
    where         rnk = 1
    order by      c.name

我正在尝试在SQL中执行查询，但未获得正确的返回值

1 个答案: