# Joins

# Joins visualized

If you are a visually oriented person, this Venn diagram may help you understand the different types of JOINs that exist within MySQL.

(opens new window)

# JOIN with subquery ("Derived" table)

SELECT x, ...
    FROM ( SELECT y, ... FROM ... ) AS a
    JOIN tbl  ON tbl.x = a.y
    WHERE ...

This will evaluate the subquery into a temp table, then JOIN that to tbl.

Prior to 5.6, there could not be an index on the temp table. So, this was potentially very inefficient:

SELECT ...
    FROM ( SELECT y, ... FROM ... ) AS a
    JOIN ( SELECT x, ... FROM ... ) AS b  ON b.x = a.y
    WHERE ...

With 5.6, the optimizer figures out the best index and creates it on the fly. (This has some overhead, so it is still not 'perfect'.)

Another common paradigm is to have a subquery to initialize something:

SELECT 
        @n := @n + 1,
        ...
    FROM ( SELECT @n := 0 ) AS initialize
    JOIN the_real_table
    ORDER BY ...

(Note: this is technically a CROSS JOIN (Cartesian product), as indicated by the lack of ON. However it is efficient because the subquery returns only one row that has to be matched to the n rows in the_real_table.)

# Full Outer Join

MySQL does not support the FULL OUTER JOIN, but there are ways to emulate one.

Setting up the data

-- ----------------------------
-- Table structure for `owners`
-- ----------------------------
DROP TABLE IF EXISTS `owners`;
CREATE TABLE `owners` (
`owner_id` int(11) NOT NULL AUTO_INCREMENT,
`owner` varchar(30) DEFAULT NULL,
PRIMARY KEY (`owner_id`)
) ENGINE=InnoDB AUTO_INCREMENT=10 DEFAULT CHARSET=latin1;

-- ----------------------------
-- Records of owners
-- ----------------------------
INSERT INTO `owners` VALUES ('1', 'Ben');
INSERT INTO `owners` VALUES ('2', 'Jim');
INSERT INTO `owners` VALUES ('3', 'Harry');
INSERT INTO `owners` VALUES ('6', 'John');
INSERT INTO `owners` VALUES ('9', 'Ellie');
-- ----------------------------
-- Table structure for `tools`
-- ----------------------------
DROP TABLE IF EXISTS `tools`;
CREATE TABLE `tools` (
`tool_id` int(11) NOT NULL AUTO_INCREMENT,
`tool` varchar(30) DEFAULT NULL,
`owner_id` int(11) DEFAULT NULL,
PRIMARY KEY (`tool_id`)
) ENGINE=InnoDB AUTO_INCREMENT=11 DEFAULT CHARSET=latin1;
-- ----------------------------
-- Records of tools
-- ----------------------------
INSERT INTO `tools` VALUES ('1', 'Hammer', '9');
INSERT INTO `tools` VALUES ('2', 'Pliers', '1');
INSERT INTO `tools` VALUES ('3', 'Knife', '1');
INSERT INTO `tools` VALUES ('4', 'Chisel', '2');
INSERT INTO `tools` VALUES ('5', 'Hacksaw', '1');
INSERT INTO `tools` VALUES ('6', 'Level', null);
INSERT INTO `tools` VALUES ('7', 'Wrench', null);
INSERT INTO `tools` VALUES ('8', 'Tape Measure', '9');
INSERT INTO `tools` VALUES ('9', 'Screwdriver', null);
INSERT INTO `tools` VALUES ('10', 'Clamp', null);

What do we want to see?

We want to get a list, in which we see who owns which tools, and which tools might not have an owner.

The queries

To accomplish this, we can combine two queries by using UNION. In this first query we are joining the tools on the owners by using a LEFT JOIN. This will add all of our owners to our resultset, doesn't matter if they actually own tools.

In the second query we are using a RIGHT JOIN to join the tools onto the owners. This way we manage to get all the tools in our resultset, if they are owned by no one their owner column will simply contain NULL. By adding a WHERE-clause which is filtering by owners.owner_id IS NULL we are defining the result as those datasets, which have not already been returned by the first query, as we are only looking for the data in the right joined table.

Since we are using UNION ALL the resultset of the second query will be attached to the first queries resultset.

SELECT `owners`.`owner`, tools.tool
FROM `owners`
LEFT JOIN `tools` ON `owners`.`owner_id` = `tools`.`owner_id`
UNION ALL
SELECT `owners`.`owner`, tools.tool
FROM `owners`
RIGHT JOIN `tools` ON `owners`.`owner_id` = `tools`.`owner_id`
WHERE `owners`.`owner_id` IS NULL;

+-------+--------------+
| owner | tool         |
+-------+--------------+
| Ben   | Pliers       |
| Ben   | Knife        |
| Ben   | Hacksaw      |
| Jim   | Chisel       |
| Harry | NULL         |
| John  | NULL         |
| Ellie | Hammer       |
| Ellie | Tape Measure |
| NULL  | Level        |
| NULL  | Wrench       |
| NULL  | Screwdriver  |
| NULL  | Clamp        |
+-------+--------------+
12 rows in set (0.00 sec)

# Joining Examples

Query to create table on db

CREATE TABLE `user` (
`id` smallint(5) unsigned NOT NULL AUTO_INCREMENT,
`name` varchar(30) NOT NULL,
`course` smallint(5) unsigned DEFAULT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB;

CREATE TABLE `course` (
`id` smallint(5) unsigned NOT NULL AUTO_INCREMENT,
`name` varchar(50) NOT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB;

Since we’re using InnoDB tables and know that user.course and course.id are related, we can specify a foreign key relationship:

ALTER TABLE `user`
ADD CONSTRAINT `FK_course`
FOREIGN KEY (`course`) REFERENCES `course` (`id`)
ON UPDATE CASCADE;

Join Query (Inner Join)

SELECT user.name, course.name
FROM `user`
INNER JOIN `course` on user.course = course.id;

# Retrieve customers with orders -- variations on a theme

This will get all the orders for all customers:

SELECT c.CustomerName, o.OrderID
    FROM Customers AS c
    INNER JOIN Orders AS o
        ON c.CustomerID = o.CustomerID
    ORDER BY c.CustomerName, o.OrderID;

This will count the number of orders for each customer:

SELECT c.CustomerName, COUNT(*) AS 'Order Count'
    FROM Customers AS c
    INNER JOIN Orders AS o
        ON c.CustomerID = o.CustomerID
    GROUP BY c.CustomerID;
    ORDER BY c.CustomerName;

Also, counts, but probably faster:

SELECT  c.CustomerName,
        ( SELECT COUNT(*) FROM Orders WHERE CustomerID = c.CustomerID ) AS 'Order Count'
    FROM Customers AS c
    ORDER BY c.CustomerName;

List only the customer with orders.

SELECT  c.CustomerName,
    FROM Customers AS c
    WHERE EXISTS ( SELECT * FROM Orders WHERE CustomerID = c.CustomerID )
    ORDER BY c.CustomerName;

# Inner-join for 3 tables

let's assume we have three table which can be used for simple website with Tags.

Fist table is for Posts.
Second for Tags
Third for Tags & Post relation

fist table "videogame"

id	title	reg_date	Content
1	BioShock Infinite	2016-08-08	....

"tags" table

id	name
1	yennefer
2	elizabeth

"tags_meta" table

post_id	tag_id
1	2

SELECT videogame.id,
    videogame.title,
    videogame.reg_date,
    tags.name,
    tags_meta.post_id
FROM tags_meta
INNER JOIN videogame ON videogame.id = tags_meta.post_id
INNER JOIN tags ON tags.id = tags_meta.tag_id
WHERE tags.name = "elizabeth"
ORDER BY videogame.reg_date

this code can return all posts which related to that tag "#elizabeth"

# Syntax

`INNER` and `OUTER` are ignored.

`FULL` is not implemented in MySQL.

"commajoin" (`FROM a,b WHERE a.x=b.y`) is frowned on; use `FROM a JOIN b ON a.x=b.y` instead.

FROM a JOIN b ON a.x=b.y includes rows that match in both tables.

FROM a LEFT JOIN b ON a.x=b.y includes all rows from `a`, plus matching data from `b`, or `NULLs` if there is no matching row.

← Error 1055: ONLYFULLGROUP_BY: something is not in GROUP BY clause ... JOINS: Join 3 table with the same name of id. →