How to check if INSERT went well in stored function?

Question

I'm creating a stored function which should insert new row to table. In this table is also one unique column.

How can I check if everything goes well and row really was inserted?

How can I check exactly that it's this unique column found (for example - try to add duplicate value)?

Answer 1

You can check the LAST_INSERT_ID() function and INSERT IGNORE.

If the INSERT IGNORE was successful, you get the primary key returned. Let's create a table with an auto increment primary key and a unique key on a name.

use test
DROP TABLE IF EXISTS nametable;
CREATE TABLE nametable
(
  id int not null auto_increment,
  name varchar(20) not null,
  primary key (id),
  unique key (name)
);
DELIMITER $$
DROP FUNCTION IF EXISTS `test`.`InsertName` $$
CREATE FUNCTION `test`.`InsertName` (newname VARCHAR(20)) RETURNS INT
BEGIN
  INSERT IGNORE INTO test.nametable (name) VALUES (newname);
  RETURN LAST_INSERT_ID();
END $$
DELIMITER ;
SELECT InsertName('rolando');
SELECT InsertName('rolando');
SELECT InsertName('pamela');
SELECT InsertName('pamela');
SHOW CREATE TABLE test.nametable\G
SELECT * FROM test.nametable;

Here is the example being run:

mysql> use test
Database changed
mysql> DROP TABLE IF EXISTS nametable;
Query OK, 0 rows affected (0.04 sec)

mysql> CREATE TABLE nametable
    -> (
    ->   id int not null auto_increment,
    ->   name varchar(20) not null,
    ->   primary key (id),
    ->   unique key (name)
    -> );
Query OK, 0 rows affected (0.07 sec)

mysql> DELIMITER $$
mysql> DROP FUNCTION IF EXISTS `test`.`InsertName` $$
Query OK, 0 rows affected (0.00 sec)

mysql> CREATE FUNCTION `test`.`InsertName` (newname VARCHAR(20)) RETURNS INT
    -> BEGIN
    ->   INSERT IGNORE INTO test.nametable (name) VALUES (newname);
    ->   RETURN LAST_INSERT_ID();
    -> END $$
Query OK, 0 rows affected (0.00 sec)

mysql> DELIMITER ;
mysql> SELECT InsertName('rolando');
+-----------------------+
| InsertName('rolando') |
+-----------------------+
|                     1 |
+-----------------------+
1 row in set (0.03 sec)

mysql> SELECT InsertName('rolando');
+-----------------------+
| InsertName('rolando') |
+-----------------------+
|                     0 |
+-----------------------+
1 row in set (0.02 sec)

mysql> SELECT InsertName('pamela');
+----------------------+
| InsertName('pamela') |
+----------------------+
|                    3 |
+----------------------+
1 row in set (0.02 sec)

mysql> SELECT InsertName('pamela');
+----------------------+
| InsertName('pamela') |
+----------------------+
|                    0 |
+----------------------+
1 row in set (0.03 sec)

mysql> SHOW CREATE TABLE test.nametable\G
*************************** 1. row ***************************
       Table: nametable
Create Table: CREATE TABLE `nametable` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `name` varchar(20) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `name` (`name`)
) ENGINE=InnoDB AUTO_INCREMENT=5 DEFAULT CHARSET=latin1
1 row in set (0.00 sec)

mysql> SELECT * FROM test.nametable;
+----+---------+
| id | name    |
+----+---------+
|  3 | pamela  |
|  1 | rolando |
+----+---------+
2 rows in set (0.00 sec)

mysql>

As shown in the preceding example, you can check the return value of the function. A nonzero return value means the INSERT IGNORE went well. A zero return value indicates a duplicate key without introducing an error number to the mysqld.

The drawback to this approach is that you cannot go back and use id 2 and 4 because of failed attempts to INSERT IGNORE in the event of a duplicate key.

Let's try another example with a different stored function setup using INSERT and without using LAST_INSERT_ID():

use test
DROP TABLE IF EXISTS nametable;
CREATE TABLE nametable
(
  id int not null auto_increment,
  name varchar(20) not null,
  primary key (id),
  unique key (name)
);
DELIMITER $$
DROP FUNCTION IF EXISTS `test`.`InsertName` $$
CREATE FUNCTION `test`.`InsertName` (newname VARCHAR(20)) RETURNS INT
BEGIN
  DECLARE rv INT;
  SELECT COUNT(1) INTO rv FROM test.nametable WHERE name = newname;
  IF rv = 0 THEN
    INSERT INTO test.nametable (name) VALUES (newname);
  END IF;
  RETURN rv;
END $$
DELIMITER ;
SELECT InsertName('rolando');
SELECT InsertName('rolando');
SELECT InsertName('pamela');
SELECT InsertName('pamela');
SHOW CREATE TABLE test.nametable\G
SELECT * FROM test.nametable;

Here is the result:

mysql> use test
Database changed
mysql> DROP TABLE IF EXISTS nametable;
Query OK, 0 rows affected, 1 warning (0.00 sec)

mysql> CREATE TABLE nametable
    -> (
    ->   id int not null auto_increment,
    ->   name varchar(20) not null,
    ->   primary key (id),
    ->   unique key (name)
    -> );
Query OK, 0 rows affected (0.10 sec)

mysql> DELIMITER $$
mysql> DROP FUNCTION IF EXISTS `test`.`InsertName` $$
Query OK, 0 rows affected (0.00 sec)

mysql> CREATE FUNCTION `test`.`InsertName` (newname VARCHAR(20)) RETURNS INT
    -> BEGIN
    ->   DECLARE rv INT;
    ->   SELECT COUNT(1) INTO rv FROM test.nametable WHERE name = newname;
    ->   IF rv = 0 THEN
    ->     INSERT INTO test.nametable (name) VALUES (newname);
    ->   END IF;
    ->   RETURN rv;
    -> END $$
Query OK, 0 rows affected (0.00 sec)

mysql> DELIMITER ;
mysql> SELECT InsertName('rolando');
+-----------------------+
| InsertName('rolando') |
+-----------------------+
|                     0 |
+-----------------------+
1 row in set (0.04 sec)

mysql> SELECT InsertName('rolando');
+-----------------------+
| InsertName('rolando') |
+-----------------------+
|                     1 |
+-----------------------+
1 row in set (0.00 sec)

mysql> SELECT InsertName('pamela');
+----------------------+
| InsertName('pamela') |
+----------------------+
|                    0 |
+----------------------+
1 row in set (0.03 sec)

mysql> SELECT InsertName('pamela');
+----------------------+
| InsertName('pamela') |
+----------------------+
|                    1 |
+----------------------+
1 row in set (0.00 sec)

mysql> SHOW CREATE TABLE test.nametable\G
*************************** 1. row ***************************
       Table: nametable
Create Table: CREATE TABLE `nametable` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `name` varchar(20) NOT NULL,
  PRIMARY KEY (`id`),
  UNIQUE KEY `name` (`name`)
) ENGINE=InnoDB AUTO_INCREMENT=3 DEFAULT CHARSET=latin1
1 row in set (0.00 sec)

mysql> SELECT * FROM test.nametable;
+----+---------+
| id | name    |
+----+---------+
|  2 | pamela  |
|  1 | rolando |
+----+---------+
2 rows in set (0.00 sec)

mysql>

In this example, the stored function returns 0 if the INSERT was OK, and returns 1 with a duplicate key on the name. The advantage? No wasted id numbers for auto_increment. The disadvantage? Doing a SELECT statement each time to check for the name already being present in the table.

You have a choice as to which way you want to handle duplicate keys. The first method lets mysqld handle the condition of the INSERT IGNORE. The second method has the stored function checking for the duplicate key first before the INSERT.