5.9.13. Running SQL Using QueryRunner

QueryRunner is a class designed to run SQL. It should be used instead of JDBC in all cases where using plain SQL is necessary and working with the ORM tools of the same purpose is not desired.

The platform’s QueryRunner is a variant of Apache DbUtils QueryRunner with the added ability to use Java Generics.

Usage example:

QueryRunner runner = new QueryRunner(persistence.getDataSource());
try {
  Set<String> scripts = runner.query("select SCRIPT_NAME from SYS_DB_CHANGELOG",
          new ResultSetHandler<Set<String>>() {
              public Set<String> handle(ResultSet rs) throws SQLException {
                  Set<String> rows = new HashSet<String>();
                  while (rs.next()) {
                      rows.add(rs.getString(1));
                  }
                  return rows;
              }
          });
  return scripts;
} catch (SQLException e) {
  throw new RuntimeException(e);
}

There are two ways of using QueryRunner: current transaction or separate transaction in autocommit mode.

  • To run a query in current transaction QueryRunner must be instantiated using a parameterless constructor. Then, query() or update() methods should be called with a Connection parameter retrieved via EntityManager.getConnection(). There is no need to close the Connection after the query, as it will be closed when the transaction is committed.

  • To run a query in a separate transaction, QueryRunner instance must be created using a constructor with the DataSource parameter retrieved using Persistence.getDataSource(). Then, query() or update() methods should be called without the Connection parameter. Connection will be created from the specified DataSource and immediately closed afterwards.