What's the difference between comma separated joins and join on syntax in MySQL?

Question

For example if I were to have a table "Person" with a column "id" that references a column "id" in table "Worker"

What would the difference between these two queries be? They yield the same results.

SELECT * 
FROM Person 
JOIN Worker 
  ON Person.id = Worker.id;

and

SELECT * 
FROM Person, 
     Worker 
WHERE Person.id = Worker.id;

Thanks

Solution 1

There is no difference at all.

First representation makes query more readable and makes it look very clear as to which join corresponds to which condition.

Solution 2

The queries are logically equivalent. The comma operator is equivalent to an [INNER] JOIN operator.

The comma is the older style join operator. The JOIN keyword was added later, and is favored because it also allows for OUTER join operations.

It also allows for the join predicates (conditions) to be separated from the WHERE clause into an ON clause. That improves (human) readability.

---

FOLLOWUP

This answer says that the two queries in the question are equivalent. We shouldn't mix old-school comma syntax for join operation with the newer JOIN keyword syntax in the same query. If we do mix them, we need to be aware of a difference in the order of precedence.

excerpt from MySQL Reference Manual

https://dev.mysql.com/doc/refman/5.6/en/join.html

INNER JOIN and , (comma) are semantically equivalent in the absence of a join condition: both produce a Cartesian product between the specified tables (that is, each and every row in the first table is joined to each and every row in the second table).

However, the precedence of the comma operator is less than that of INNER JOIN, CROSS JOIN, LEFT JOIN, and so on. If you mix comma joins with the other join types when there is a join condition, an error of the form Unknown column 'col_name' in 'on clause' may occur. Information about dealing with this problem is given later in this section.

Solution 3

Beside better readability, there is one more case where explicitly joined tables are better instead of comma-separated tables.

let's see an example:

Create Table table1
(
    ID int NOT NULL Identity(1, 1) PRIMARY KEY ,
    Name varchar(50)
)

Create Table table2
(
    ID int NOT NULL Identity(1, 1) PRIMARY KEY ,
    ID_Table1 INT NOT NULL
)

Following query will give me all columns and rows from both tables

SELECT
    *
FROM table1, table2

Following query will give me columns from first table with table alias called 'table2'

SELECT
    *
FROM table1 table2

If you mistakenly forget comma in comma-separated join, second table automatically convert to table alias for first table. Not in all cases, but there is chances for something like this

Solution 4

Using JOINS makes the code easier to read, since it's self-explanatory.

In speed there is no difference (I tested it) and the execution plan is the same

If the query optimizer is doing its job right, there should be no difference between those queries. They are just two ways to specify the same desired result.

Solution 5

The SELECT * FROM table1, table2, etc. is good for a couple of tables, but it becomes exponentially harder as the number of tables increases.

The JOIN syntax makes it explicit what criteria affects which tables (giving a condition). Also, the second way is the older standard.

Although, to the database, they end up being the same

View more solutions

Comments

[Liam]

Is there any performance difference? How about in PostgreSQL?

[Sateesh Pagolu]

There is absolutely no difference in performance. Final version prepared by SQL Server would be the same in both the cases.

[Gal]

exponentially harder? how? and for who?

[wobbily_col]

For larger queries I would say that the first way is more readable. If you have a mix left joins and some inner joins, then it is far more consistent to have the joins condition on the join clause, rather than some in the joins and some in the where clause.

[Stephen York]

The first is way more readable and explicit

[Captain Payalytic]

This answer is not correct. In the MySQL manual on the page dev.mysql.com/doc/refman/5.5/en/join.html it states: "However, the precedence of the comma operator is less than that of INNER JOIN, CROSS JOIN, LEFT JOIN, and so on. If you mix comma joins with the other join types when there is a join condition, an error of the form Unknown column 'col_name' in 'on clause' may occur"

[spencer7593]

@CaptainPayalytic: The two queries given in the question are logically equivalent. As you point out, there is actually a difference between the two notations. (I concur with you that this answer is wrong. There is, in fact, a difference between the two queries... we could say that one uses the old school comma syntax for the join operation, the other uses the JOIN keyword.)

[philipxy]

comma is CROSS JOIN or INNER JOIN ON 1=1. It has lower precedence than keyword joins. INNER JOIN ON is defined in terms of comma & WHERE.

[youcantryreachingme]

Further to earlier comments, using the first format you can group all join conditions into ON clauses, leaving your WHERE clauses to specify filters for your data (with the caveat that it's possible to write join conditions that act as filters, but in general, keeping to this practice makes the code more readable).

[Matt Arnold]

If table1 and table2 are unrelated, I would rate the FROM + comma syntax here above the alternative and more verbose INNER JOIN table2 ON 1 = 1 since the ON predicate is redundant.

[spencer7593]

@MattArnold : with MySQL, the ON clause is optional. the ON 1=1 has no effect and is unnecessary. Also, the INNER keyword is optional, and has no effect. i.e. FROM table1 JOIN table2 is sufficient, although in this case, with no join condition in the ON or WHERE clause, we could include the optional CROSS keyword as an aid to the future reader. To help the reader comprehend that the absence of a join condition was intentional. i.e. FROM table1 CROSS JOIN table2. I think this form communicates our intent more clearly than the comma syntax.

[Raymond Nijland]

Run a EXPLAIN <query>/SHOW WARNINGS and you will see those queries optimizes/internally rewriten (notice the warning message) the same -> db-fiddle.com/f/us9ffeFw7vcMVt2WoNXQzm/0

[Sнаđошƒаӽ]

@Gal I think it is just an indian way of saying that comma separated table names can be written in a single line... ¯\_(ツ)_/¯ That's all I can make out of that statement, coz I am only guessing - only the author can know the meaning of that..

[amy]

Are they also the same in Oracle?

[Jeremy Ray Brown]

Easier to read is a great explanation. Code should always be written so it is easier to read. I agree with you 100%