Concurrency issues when retriveing Ids of newly inserted rows with ibatis

Question

I'm using iBatis/Java and Postgres 8.3. When I do an insert in ibatis i need the id returned.
I use the following table for describing my question:
CREATE TABLE sometable ( id serial NOT NULL, somefield VARCHAR(10) );
The Sequence sometable_id_seq gets autogenerated by running the create statement.

At the moment i use the following sql map:

<insert id="insertValue" parameterClass="string" >
 INSERT INTO sometable ( somefield ) VALUES ( #value# );
 <selectKey keyProperty="id" resultClass="int">
  SELECT last_value AS id FROM sometable_id_seq
 </selectKey>
</insert>

It seems this is the ibatis way of retrieving the newly inserted id. Ibatis first runs a INSERT statement and afterwards it asks the sequence for the last id.
I have doubts that this will work with many concurrent inserts.

Could this cause problems? Like returning the id of the wrong insert?

( See also my related question about how to get ibatis to use the INSERT .. RETURING .. statements )

Answer 1

This is definitely wrong. Use:

select currval('sometable_id_seq')

or better yet:

INSERT INTO sometable ( somefield ) VALUES ( #value# ) returning id

which will return you inserted id.

Answer 2

Here is simple example:

<statement id="addObject"
        parameterClass="test.Object"
        resultClass="int">
        INSERT INTO objects(expression, meta, title,
        usersid)
        VALUES (#expression#, #meta#, #title#, #usersId#)
        RETURNING id
</statement>

And in Java code:

Integer id = (Integer) executor.queryForObject("addObject", object);
object.setId(id);

Answer 3

I have another thought. ibatis invokes the insert method delegate the Class: com.ibatis.sqlmap.engine.impl.SqlMapExecutorDelegate,with the code:

 try {
      trans = autoStartTransaction(sessionScope, autoStart, trans);

      SelectKeyStatement selectKeyStatement = null;
      if (ms instanceof InsertStatement) {
        selectKeyStatement = ((InsertStatement) ms).getSelectKeyStatement();
      }

      // Here we get the old value for the key property. We'll want it later if for some reason the
      // insert fails.
      Object oldKeyValue = null;
      String keyProperty = null;
      boolean resetKeyValueOnFailure = false;
      if (selectKeyStatement != null && !selectKeyStatement.isRunAfterSQL()) {
        keyProperty = selectKeyStatement.getKeyProperty();
        oldKeyValue = PROBE.getObject(param, keyProperty);
        generatedKey = executeSelectKey(sessionScope, trans, ms, param);
        resetKeyValueOnFailure = true;
      }

      StatementScope statementScope = beginStatementScope(sessionScope, ms);
      try {
        ms.executeUpdate(statementScope, trans, param);
      }catch (SQLException e){
        // uh-oh, the insert failed, so if we set the reset flag earlier, we'll put the old value
        // back...
        if(resetKeyValueOnFailure) PROBE.setObject(param, keyProperty, oldKeyValue);
        // ...and still throw the exception.
        throw e;
      } finally {
        endStatementScope(statementScope);
      }

      if (selectKeyStatement != null && selectKeyStatement.isRunAfterSQL()) {
        generatedKey = executeSelectKey(sessionScope, trans, ms, param);
      }

      autoCommitTransaction(sessionScope, autoStart);
    } finally {
      autoEndTransaction(sessionScope, autoStart);
    }

You can see that the insert and select operator are in a Transaction. So I think there is no concureency problem with the insert method.