问题
In Hibernate, you can use @SQLInsert to define a custom insert query. Hibernate is using prepared statements, so you just need to provide the ? as placeholder, like this: INSERT INTO table (colA, colB) VALUES (?,?)
On a duplicate key for colA, I want colB to be updated with the new value, so i tried INSERT INTO table (colA, colB) VALUES (?,?) ON DUPLICATE KEY UPDATE colB = ? - This however throws the error that no third parameter is given. (No value specified for parameter 3)
How is the correct way to write this Query for hibernate? Any non-data-dependent update like INSERT INTO table (colA, colB) VALUES (?,?) ON DUPLICATE KEY UPDATE colB = colB +1 is working - but I need to set the actual value of colB that has been passed with the insert-call.
回答1:
The correct answer would be to use the following statement:
INSERT INTO table (colA, colB) VALUES (?,?) ON DUPLICATE KEY
UPDATE colB = VALUES(colB);
However then the problem appears that hibernate did not receive an auto-increment value back, once the update statement has been performed.
I found the following blog post (http://www.jroller.com/mmatthews/entry/getting_hibernate_and_mysql_s) and modifed the query to the following:
INSERT INTO table (colA, colB) VALUES (?,?) ON DUPLICATE KEY
UPDATE colB = VALUES(colB), id = LAST_INSERT_ID(id);
which finally works.
An unsolvable Problem with this approach is, that the insertion of two entities that are equal within the same transaction does not work. Even if the second insertion would cause the correct update, the em would end up with 2 entity-instances representing the same database row - which is not allowed.
To solve this, one just needs to ensure that you don't insert 2 entities which are rendered equal due to their constraints. (I used the same logic for equals/hashcode as the composite-unique-key constraint, so i am able to eliminate such duplicates when performing batch-inserts)
来源：https://stackoverflow.com/questions/28580151/hibernate-sqlinsert-and-on-duplicate-key