MySQL Error Handling in Stored Procedures

출처 : http://www.mysqltutorial.org/mysql-error-handling-in-stored-procedures/

Summary: in this tutorial, you will learn how to use MySQL handler to handle exceptions or errors encountered in stored procedures.

When an error occurs inside a stored procedure, it is important to handle it appropriately, such as continuing or exiting the current code block’s execution, and issuing a meaningful error message.

MySQL provides an easy way to define handlers that handle from general conditions such as warnings or exceptions to specific conditions e.g., specific error codes.

Declaring a handler

To declare a handler, you use the  DECLARE HANDLER statement as follows:

1	DECLARE action HANDLER FOR condition_value statement;

If a condition whose value matches the  condition_value , MySQL will execute the statement and continue or exit the current code block based on the action .

The action accepts one of the following values:

• CONTINUE :  the execution of the enclosing code block ( BEGIN … END ) continues.
• EXIT : the execution of the enclosing code block, where the handler is declared, terminates.

The  condition_value specifies a particular condition or a class of conditions that activate the handler. The  condition_value accepts one of the following values:

• A MySQL error code.
• A standard SQLSTATE value. Or it can be an SQLWARNING , NOTFOUND or SQLEXCEPTION condition, which is shorthand for the class of SQLSTATE values. The NOTFOUND condition is used for a cursor or  SELECT INTO variable_list statement.
• A named condition associated with either a MySQL error code or SQLSTATE value.

The statement could be a simple statement or a compound statement enclosing by the BEGIN and ENDkeywords.

MySQL error handling examples

Let’s look into several examples of declaring handlers.

The following handler means that if an error occurs, set the value of the  has_error variable to 1 and continue the execution.

1	DECLARE CONTINUE HANDLER FOR SQLEXCEPTION SET has_error = 1;

The following is another handler which means that in case an error occurs, rollback the previous operation, issue an error message, and exit the current code block. If you declare it inside the BEGIN ENDblock of a stored procedure, it will terminate stored procedure immediately.

1 2 3 4 5

DECLARE EXIT HANDLER FOR SQLEXCEPTION BEGIN ROLLBACK; SELECT 'An error has occurred, operation rollbacked and the stored procedure was terminated'; END;

The following handler means that if there are no more rows to fetch, in case of a cursor or SELECT INTO statement, set the value of the  no_row_found variable to 1 and continue execution.

1	DECLARE CONTINUE HANDLER FOR NOT FOUND SET no_row_found = 1;

The following handler means that if a duplicate key error occurs, MySQL error 1062 is issued. It issues an error message and continues execution.

1 2	DECLARE CONTINUE HANDLER FOR 1062 SELECT 'Error, duplicate key occurred';

MySQL handler example in stored procedures

First, we create a new table named  article_tags for the demonstration:

1 2 3 4 5

CREATE TABLE article_tags(     article_id INT,     tag_id     INT,     PRIMARY KEY(article_id,tag_id) );

The  article_tags table stores the relationships between articles and tags. Each article may have many tags and vice versa. For the sake of simplicity, we don’t create articles and tags tables, as well as the foreign keys in the  article_tags table.

Next, we create a stored procedure that inserts article id and tag id into the article_tags table:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

DELIMITER $$   CREATE PROCEDURE insert_article_tags(IN article_id INT, IN tag_id INT) BEGIN   DECLARE CONTINUE HANDLER FOR 1062 SELECT CONCAT('duplicate keys (',article_id,',',tag_id,') found') AS msg;   -- insert a new record into article_tags INSERT INTO article_tags(article_id,tag_id) VALUES(article_id,tag_id);   -- return tag count for the article SELECT COUNT(*) FROM article_tags; END

Then, we add tag id 1, 2 and 3 for the article 1 by calling the insert_article_tags  stored procedure as follows:

1 2 3

CALL insert_article_tags(1,1); CALL insert_article_tags(1,2); CALL insert_article_tags(1,3);

After that, we try to insert a duplicate key to check if the handler is really invoked.

1	CALL insert_article_tags(1,3);

We got an error message. However, because we declared the handler as a CONTINUE handler, the stored procedure continued the execution. As the result, we got the tag count for the article as well.

If we change the CONTINUE in the handler declaration to EXIT , we will get an error message only.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

DELIMITER $$   CREATE PROCEDURE insert_article_tags_2(IN article_id INT, IN tag_id INT) BEGIN   DECLARE EXIT HANDLER FOR SQLEXCEPTION SELECT 'SQLException invoked';   DECLARE EXIT HANDLER FOR 1062         SELECT 'MySQL error code 1062 invoked';   DECLARE EXIT HANDLER FOR SQLSTATE '23000' SELECT 'SQLSTATE 23000 invoked';   -- insert a new record into article_tags INSERT INTO article_tags(article_id,tag_id)    VALUES(article_id,tag_id);   -- return tag count for the article SELECT COUNT(*) FROM article_tags; END

Finally, we can try to add a duplicate key to see the effect.

1	CALL insert_article_tags_2(1,3);

MySQL handler precedence

In case there are multiple handlers that are eligible for handling an error, MySQL will call the most specific handler to handle the error first.

An error always maps to one MySQL error code because in MySQL it is the most specific. An SQLSTATEmay map to many MySQL error codes, therefore, it is less specific. An SQLEXCPETION or an SQLWARNING is the shorthand for a class of SQLSTATES values so it is the most generic.

Based on the handler precedence’s rules,  MySQL error code handler, SQLSTATE handler and SQLEXCEPTIONtakes the first, second and third precedence.

Suppose we declare three handlers in the  insert_article_tags_3 stored procedure as follows:

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

DELIMITER $$   CREATE PROCEDURE insert_article_tags_3(IN article_id INT, IN tag_id INT) BEGIN   DECLARE EXIT HANDLER FOR 1062 SELECT 'Duplicate keys error encountered'; DECLARE EXIT HANDLER FOR SQLEXCEPTION SELECT 'SQLException encountered'; DECLARE EXIT HANDLER FOR SQLSTATE '23000' SELECT 'SQLSTATE 23000';   -- insert a new record into article_tags INSERT INTO article_tags(article_id,tag_id) VALUES(article_id,tag_id);   -- return tag count for the article SELECT COUNT(*) FROM article_tags; END

We try to insert a duplicate key into the article_tags table by calling the stored procedure:

1	CALL insert_article_tags_3(1,3);

As you see the MySQL error code handler is called.

Using a named error condition

Let’s start with an error handler declaration.

1 2	DECLARE EXIT HANDLER FOR 1051 SELECT 'Please create table abc first'; SELECT * FROM abc;

What does the number 1051 really mean? Imagine you have a big stored procedure polluted with those numbers all over places; it will become a nightmare to maintain the code.

Fortunately, MySQL provides us with the DECLARE CONDITION statement that declares a named error condition, which associates with a condition.

The syntax of the DECLARE CONDITION statement is as follows:

1	DECLARE condition_name CONDITION FOR condition_value;

The condition_value  can be a MySQL error code such as 1015 or a SQLSTATE value. The condition_valueis represented by the condition_name .

After the declaration, we can refer to condition_name instead of condition_value .

So we can rewrite the code above as follows:

1 2 3 4 5 6

DECLARE table_not_found CONDITION for 1051;   DECLARE EXIT HANDLER FOR table_not_found SELECT 'Please create table abc first';   SELECT * FROM abc;

This code is obviously more readable than the previous one.

Notice that the condition declaration must appear before handler or cursor declarations.