MySQL UNIQUE declaration

Question

What's the difference between this two declaration:

CREATE TABLE `t1` (`col1` INT, `col2` INT, UNIQUE(`col1`,`col2`));

or

CREATE TABLE `t2` (`col1` INT UNIQUE, `col2` INT UNIQUE);

Answer 1

You've used two different ways of specifying keys, but your keys are actually different.

• The first requires that the combination of col1 and col2 be unique. So you couldn't have (1,2)(1,2)(1,2) but you could have (1,2)(1,3)(2,3) .

• The second requires that each column be unique, independently; in that case, even the latter example is invalid.

Here's the first statement, with the same syntactic approach but rewritten to match the semantics of the second:

CREATE TABLE `t1` (`col1` INT, `col2` INT, UNIQUE(`col1`), UNIQUE(`col2`));

You can read more about the syntax of table schemas in the documentation .

Answer 2

If you write :

`CREATE TABLE `t1` (`col1` INT, `col2` INT, UNIQUE(`col1`,`col2`));`

That means you can't have the same duo of col1 + col2 so :

col1 | col2
1    | 1
1    | 2
2    | 1
2    | 2

It's ok but :

col1 | col2
1    | 1
1    | 2
2    | 1
2    | 2
1    | 1  // NOT OK, duplicate of first row
2    | 2  // NOT OK, duplicate of row number 4

Now if you write :

CREATE TABLE `t2` (`col1` INT UNIQUE, `col2` INT UNIQUE);

That means you can't have two identical value for col1 , neither for col2 . So :

col1 | col2
1    | 1   // OK
1    | 2   // NOT OK, you already have value 1 for col1 in row 1
3    | 1   // NOT OK, you already have value 1 for col2 in row 1
4    | 3   // OK

Is it clear for you?