How to join all three tables in MYSQL?

Question

I have a MySQL database that has three tables holding information about uploaded photos by users. I have a PHP page that displays all the photos in the database ( tbl_uploads ) and who uploaded them ( tbl_users ). To display the photos and who uploaded them I have a join in the MySQL query.

SELECT * 
FROM tbl_uploads, tbl_users
WHERE tbl_uploads.user_id = tbl_users.user_id
ORDER BY date DESC

I now want to join a third table tbl_collab to the MySQL query that allows me to display all the users that collaborated with the photo (a form allows them to post the $file and their $ user_id to tbl_collab ). I guess I need to add a join from tbl_uploads.file with tbl_collab.user_id but I'm not sure how.

tbl_users
|//**user_id**//|int(11)|No|
|user_name|varchar(25)|No|
|user_email|varchar(60)|No|
|user_password|varchar(255)|No|
|joining_date|datetime|No|

tbl_uploads

|//**id**//|int(10)|No|
|file|varchar(100)|No|
|type|varchar(30)|No|
|size|int(11)|No|
|user_id|int(11)|No|
|user_name|varchar(25)|No|

 tbl_collab

|//**id**//|int(11)|No|
|user_name|varchar(100)|No|
|user_id|int(11)|No|
|file|varchar(255)|No|

I have been trying your various suggestions and I can't really get them to work as I would hope so I have made a mysql fiddle that might be help me. The problem is that when I loop through the rows that the query throws up in PHP I ether get just the rows where there is join with tbl_uploads.file and tbl.collab.file or I get the multiple rows duplicating themselves.

Answer 1

I'd suggest preferring ANSI SQL syntax for joins (over mentioning multiple tables in the "from" clause) as once the queries get complex I find the ANSI syntax easier to follow. Using that syntax, joining multiple tables is no big deal. eg,

SELECT uploads.<column>, users.<column>, collabs.<column> 
FROM tbl_uploads uploads
JOIN tbl_users users ON users.user_id=uploads.user_id
JOIN tbl_collabs collabs ON collabs.file=uploads.file
ORDER BY uploads.date DESC

(Note, replace <column> above with the names of columns you want to select from the respective tables, using AS syntax to provide unique names where necessary.)

Consider that you will probably want to create indexes over the fields in the join conditions for performance if you expect the database will become large. You may also want to use left joins when joining, eg, tbl_collabs if it is possible an upload will have no collaborators, otherwise the query will return no data if there are no matching rows in tbl_collabs .

Answer 2

Seems the join is on file to me

  SELECT * 
  FROM tbl_uploads
  inner join tbl_users on  tbl_uploads.user_id = tbl_users.user_id
  inner join tbl_collab  on tbl_collab.file = tbl_uploads.file
  ORDER BY date DESC

Answer 3

You can just add another join condition. Also, note that implicit joins (having multiple tables in the from clause) isn't considered a good practice, and you should probably use explicit join clauses:

SELECT   * 
FROM     tbl_uploads up
JOIN     tbl_users us ON up.user_id = us.user_id
JOIN     tbl_collab c ON c.user_id = up.user_id
ORDER BY date DESC

Answer 4

The first thing to do is to normalize your data. If you look closely, username appears in all three tables. It shouldn't. It belongs only in the users table. Then your other tables need to have a user_id field instead of the username.

tbl_uploads

|//**id**//|int(10)|No|
|file|varchar(100)|No|
|type|varchar(30)|No|
|size|int(11)|No|
|user_id|int(11)|No|

 tbl_collab

|//**id**//|int(11)|No|
|user_id|int(11)|No|
|file|varchar(255)|No|

In both cases the user_id is a foreign key to the id field in the users table. Now we have something consistent to join on.

SELECT * FROM tbl_uploads
  INNER JOIN tbl_users ON tbl_uploads.user_id = tbl_users.user_id
  INNER JOIN tbl_collab ON tbl_collab.file = tbl_uploads.file

Whether you should use INNER JOIN or LEFT JOIN depends on exactly what you need to do with your data, but INNER JOIN seems more appropriate based on information provided.

Update : As @drew pointed out, none of your tables have a column named date did you perhaps intend to sort by tbl_users.joining_date ?