[mysql] MySQL JOIN ON vs USING?

In a MySQL JOIN, what is the difference between ON and USING()? As far as I can tell, USING() is just more convenient syntax, whereas ON allows a little more flexibility when the column names are not identical. However, that difference is so minor, you'd think they'd just do away with USING().

Is there more to this than meets the eye? If yes, which should I use in a given situation?

For those experimenting with this in phpMyAdmin, just a word:

phpMyAdmin appears to have a few problems with USING. For the record this is phpMyAdmin run on Linux Mint, version: "4.5.4.1deb2ubuntu2", Database server: "10.2.14-MariaDB-10.2.14+maria~xenial - mariadb.org binary distribution".

I have run SELECT commands using JOIN and USING in both phpMyAdmin and in Terminal (command line), and the ones in phpMyAdmin produce some baffling responses:

1) a LIMIT clause at the end appears to be ignored.
2) the supposed number of rows as reported at the top of the page with the results is sometimes wrong: for example 4 are returned, but at the top it says "Showing rows 0 - 24 (2503 total, Query took 0.0018 seconds.)"

Logging on to mysql normally and running the same queries does not produce these errors. Nor do these errors occur when running the same query in phpMyAdmin using JOIN ... ON .... Presumably a phpMyAdmin bug.

Thought I would chip in here with when I have found ON to be more useful than USING. It is when OUTER joins are introduced into queries.

ON benefits from allowing the results set of the table that a query is OUTER joining onto to be restricted while maintaining the OUTER join. Attempting to restrict the results set through specifying a WHERE clause will, effectively, change the OUTER join into an INNER join.

Granted this may be a relative corner case. Worth putting out there though.....

For example:

CREATE TABLE country (
   countryId int(10) unsigned NOT NULL PRIMARY KEY AUTO_INCREMENT,
   country varchar(50) not null,
  UNIQUE KEY countryUIdx1 (country)
) ENGINE=InnoDB;

insert into country(country) values ("France");
insert into country(country) values ("China");
insert into country(country) values ("USA");
insert into country(country) values ("Italy");
insert into country(country) values ("UK");
insert into country(country) values ("Monaco");


CREATE TABLE city (
  cityId int(10) unsigned NOT NULL PRIMARY KEY AUTO_INCREMENT,
  countryId int(10) unsigned not null,
  city varchar(50) not null,
  hasAirport boolean not null default true,
  UNIQUE KEY cityUIdx1 (countryId,city),
  CONSTRAINT city_country_fk1 FOREIGN KEY (countryId) REFERENCES country (countryId)
) ENGINE=InnoDB;


insert into city (countryId,city,hasAirport) values (1,"Paris",true);
insert into city (countryId,city,hasAirport) values (2,"Bejing",true);
insert into city (countryId,city,hasAirport) values (3,"New York",true);
insert into city (countryId,city,hasAirport) values (4,"Napoli",true);
insert into city (countryId,city,hasAirport) values (5,"Manchester",true);
insert into city (countryId,city,hasAirport) values (5,"Birmingham",false);
insert into city (countryId,city,hasAirport) values (3,"Cincinatti",false);
insert into city (countryId,city,hasAirport) values (6,"Monaco",false);

-- Gah. Left outer join is now effectively an inner join 
-- because of the where predicate
select *
from country left join city using (countryId)
where hasAirport
; 

-- Hooray! I can see Monaco again thanks to 
-- moving my predicate into the ON
select *
from country co left join city ci on (co.countryId=ci.countryId and ci.hasAirport)
; 

Wikipedia has the following information about USING:

The USING construct is more than mere syntactic sugar, however, since the result set differs from the result set of the version with the explicit predicate. Specifically, any columns mentioned in the USING list will appear only once, with an unqualified name, rather than once for each table in the join. In the case above, there will be a single DepartmentID column and no employee.DepartmentID or department.DepartmentID.

Tables that it was talking about:

The Postgres documentation also defines them pretty well:

The ON clause is the most general kind of join condition: it takes a Boolean value expression of the same kind as is used in a WHERE clause. A pair of rows from T1 and T2 match if the ON expression evaluates to true.

The USING clause is a shorthand that allows you to take advantage of the specific situation where both sides of the join use the same name for the joining column(s). It takes a comma-separated list of the shared column names and forms a join condition that includes an equality comparison for each one. For example, joining T1 and T2 with USING (a, b) produces the join condition ON T1.a = T2.a AND T1.b = T2.b.

Furthermore, the output of JOIN USING suppresses redundant columns: there is no need to print both of the matched columns, since they must have equal values. While JOIN ON produces all columns from T1 followed by all columns from T2, JOIN USING produces one output column for each of the listed column pairs (in the listed order), followed by any remaining columns from T1, followed by any remaining columns from T2.

Database tables

To demonstrate how the USING and ON clauses work, let's assume we have the following post and post_comment database tables, which form a one-to-many table relationship via the post_id Foreign Key column in the post_comment table referencing the post_id Primary Key column in the post table:

The parent post table has 3 rows:

| post_id | title     |
|---------|-----------|
| 1       | Java      |
| 2       | Hibernate |
| 3       | JPA       |

and the post_comment child table has the 3 records:

| post_comment_id | review    | post_id |
|-----------------|-----------|---------|
| 1               | Good      | 1       |
| 2               | Excellent | 1       |
| 3               | Awesome   | 2       |

The JOIN ON clause using a custom projection

Traditionally, when writing an INNER JOIN or LEFT JOIN query, we happen to use the ON clause to define the join condition.

For example, to get the comments along with their associated post title and identifier, we can use the following SQL projection query:

SELECT
   post.post_id,
   title,
   review
FROM post
INNER JOIN post_comment ON post.post_id = post_comment.post_id
ORDER BY post.post_id, post_comment_id

And, we get back the following result set:

| post_id | title     | review    |
|---------|-----------|-----------|
| 1       | Java      | Good      |
| 1       | Java      | Excellent |
| 2       | Hibernate | Awesome   |

The JOIN USING clause using a custom projection

When the Foreign Key column and the column it references have the same name, we can use the USING clause, like in the following example:

SELECT
  post_id,
  title,
  review
FROM post
INNER JOIN post_comment USING(post_id)
ORDER BY post_id, post_comment_id

And, the result set for this particular query is identical to the previous SQL query that used the ON clause:

| post_id | title     | review    |
|---------|-----------|-----------|
| 1       | Java      | Good      |
| 1       | Java      | Excellent |
| 2       | Hibernate | Awesome   |

The USING clause works for Oracle, PostgreSQL, MySQL, and MariaDB. SQL Server doesn't support the USING clause, so you need to use the ON clause instead.

The USING clause can be used with INNER, LEFT, RIGHT, and FULL JOIN statements.

SQL JOIN ON clause with SELECT *

Now, if we change the previous ON clause query to select all columns using SELECT *:

SELECT *
FROM post
INNER JOIN post_comment ON post.post_id = post_comment.post_id
ORDER BY post.post_id, post_comment_id

We are going to get the following result set:

| post_id | title     | post_comment_id | review    | post_id |
|---------|-----------|-----------------|-----------|---------|
| 1       | Java      | 1               | Good      | 1       |
| 1       | Java      | 2               | Excellent | 1       |
| 2       | Hibernate | 3               | Awesome   | 2       |

As you can see, the post_id is duplicated because both the post and post_comment tables contain a post_id column.

SQL JOIN USING clause with SELECT *

On the other hand, if we run a SELECT * query that features the USING clause for the JOIN condition:

SELECT *
FROM post
INNER JOIN post_comment USING(post_id)
ORDER BY post_id, post_comment_id

We will get the following result set:

| post_id | title     | post_comment_id | review    |
|---------|-----------|-----------------|-----------|
| 1       | Java      | 1               | Good      |
| 1       | Java      | 2               | Excellent |
| 2       | Hibernate | 3               | Awesome   |

You can see that this time, the post_id column is deduplicated, so there is a single post_id column being included in the result set.

Conclusion

If the database schema is designed so that Foreign Key column names match the columns they reference, and the JOIN conditions only check if the Foreign Key column value is equal to the value of its mirroring column in the other table, then you can employ the USING clause.

Otherwise, if the Foreign Key column name differs from the referencing column or you want to include a more complex join condition, then you should use the ON clause instead.