Reputation: 9829
I have a database schema like this picture:
I want to write a query that select data of all 6 tables and a field that indicate whether a specific user applied for a job or not.
I've tried:
SELECT j.id, j.expired_date, j.title, j.views, j.status
, j.unread, j.applicants, j.location, j.created_date
, j.contract_type, c.country
, group_concat(DISTINCT jp.page_name) AS fan_pages
, group_concat(DISTINCT jp.id_page) AS id_pages
, app.id AS applied
FROM jobs AS j
INNER JOIN country AS c ON c.id = j.country
LEFT JOIN job_pages AS jp ON jp.id_job = j.id
LEFT JOIN applications AS app ON app.id_job = j.id
LEFT JOIN resumes AS res ON res.id = app.id_resume
LEFT JOIN applicants AS alc ON alc.id = res.id_applicant
AND alc.id_user = 15
WHERE ( j.status = 0 )
AND ( j.expired_date = 0
OR j.expired_date > 1323228856 )
GROUP BY `j`.`id`
ORDER BY `j`.`id` desc
LIMIT 5 OFFSET 5
But it return a result that indicates a job was applied by any user. How can I rewrite this query?
Upvotes: 0
Views: 755
Reputation: 19999
-- Edit --
Below is a basic ERD for how it would be easier to track users who have applied to jobs. I made the relationship between User and Resume a 1:M, in case you wanted to track resume versions. If not, it should be a 1:1 relationship.
So given the ERD, you have a user apply to a job with a resume. If you want to make the resume optional, then you remove the Resume table from the M:M with Job and link directly to User.
Just some ideas...
-- Original --
Just some advice.
Seems to me that you may need to re-visit the schema design. It seems like the applicants table should be a pivot table between the users and jobs tables. The users and jobs table have a M:M relationship in that many users can apply to many jobs and many jobs can be applied to by many users. The applicants table should act as a transactional table when a user applies to a job.
Also, shouldn't the resumes table be directly linked to the users table? How can an application own a resume?
User owns a resume.
User applies to a job with a resume (applicant).
Upvotes: 2
Reputation: 3033
Try it,
SELECT j.id_user as creator, alc.id_user as applier, j.id , j.expired_date, j.title, j.views, j.status
, j.unread, j.applicants, j.location, j.created_date
, j.contract_type, c.country
, group_concat(DISTINCT jp.page_name) AS fan_pages
, group_concat(DISTINCT jp.id_page) AS id_pages
, MAX(app.id) AS applied
FROM jobs AS j
INNER JOIN country AS c ON c.id = j.country
LEFT JOIN job_pages AS jp ON jp.id_job = j.id
LEFT JOIN applications AS app ON app.id_job = j.id
LEFT JOIN resumes AS res ON res.id = app.id_resume
LEFT JOIN applicants AS alc ON alc.id = res.id_applicant
WHERE
( alc.id_user = 15 or alc.id_user IS NULL) AND
( j.status = 0 )
AND ( j.expired_date = 0
OR j.expired_date > 1323228856 )
GROUP BY `j`.`id`
ORDER BY `j`.`id` desc
I believe that, now the query is better:
SELECT
j.id, j.expired_date, j.title, j.views, j.status
, j.unread, j.applicants, j.location, j.created_date
, j.contract_type, c.country
, group_concat(DISTINCT jp.page_name) AS fan_pages
, group_concat(DISTINCT jp.id_page) AS id_pages
, max(app.id) AS applied
FROM users AS u
LEFT JOIN jobs AS j ON 1
INNER JOIN country AS c ON c.id = j.country
LEFT JOIN job_pages AS jp ON jp.id_job = j.id
LEFT JOIN applicants AS alc ON alc.id_user = u.id
LEFT JOIN resumes AS res ON res.id_applicant = alc.id
LEFT JOIN applications AS app ON app.id_resume = res.id AND app.id_job = j.id
WHERE u.id = 16 AND
( j.status = 0 )
AND ( j.expired_date = 0 OR j.expired_date > 1323228856 )
GROUP BY j.id
ORDER BY j.id
New updates:
Upvotes: 1