Spring JDBC Framework详解——批量JDBC操作、ORM映射

作者:niumd,转载请注明出处,谢谢

发表时间:2010 年 03 月 17 日

一、Spring JDBC 概述

Spring 提供了一个强有力的模板类JdbcTemplate简化JDBC操作,DataSource,JdbcTemplate都可以以Bean的方式定义在想xml配置文件,JdbcTemplate创建只需注入一个DataSource,应用程序Dao层只需要继承JdbcDaoSupport, 或者注入JdbcTemplate,便可以获取JdbcTemplate,JdbcTemplate是一个线程安全的类,多个Dao可以注入一个JdbcTemplate;

public class JdbcCustomerDao extends JdbcDaoSupport implements CustomerDao { 
} 



public class JdbcCustomerDao implements CustomerDao { 




private JdbcTemplate jdbcTemplate 




public void setJdbcTemplate()JdbcTemplate jdbcTemplate{ 


this.jdbcTemplate=jdbcTemplate 

} 
}
public class JdbcCustomerDao extends JdbcDaoSupport implements CustomerDao {
}

public class JdbcCustomerDao implements CustomerDao {

          private JdbcTemplate jdbcTemplate

         public void setJdbcTemplate()JdbcTemplate jdbcTemplate{
               this.jdbcTemplate=jdbcTemplate
         }
}

二、 JdbcTemplate 提供以下主要方法简化JDBC操作:

2.1、List query(String sql,Ojbect[] args,RowMapper rowMapper)

说明:常用的查询,sql待执行的sql语句,args是sql语句的参数,rowMapper负责将每一行记录转化为java对象存放在list,并最终返回,例如:

public List<Book> queryByAuthor(String author) { 

String sql = "select * from book where author=?"; 

Collection c = getJdoTemplate().find(sql, 

new Object[] { author },new BookRowMapper()); 


List<Book> books = new ArrayList<Book>(); 

books.addAll(c); 

return books; 

} 



class BookRowMapper implements RowMapper{ 


public Object mapRow(ResultSet res, int index) throws SQLException { 


Book book = new Book(); 


book.setId(rs.getInt("id")); 

//省略set 

return book; 

} 
}
public List<Book> queryByAuthor(String author) {
        String sql = "select * from book where author=?";
        Collection c = getJdoTemplate().find(sql,
                new Object[] { author },new BookRowMapper());
        List<Book> books = new ArrayList<Book>();
        books.addAll(c);
        return books;
}

class BookRowMapper implements RowMapper{
     public Object mapRow(ResultSet res, int index) throws SQLException {
          Book book = new Book();
          book.setId(rs.getInt("id"));
          //省略set
       return book;
    }
}

更新、删除、其他查询操作类似,举例如下,详细细节请参考spring api:

//返回值为一个长整形 

public long getAverageAge() { 


return getJdbcTemplate().queryForLong("SELECT AVG(age) FROM employee"); 

} 
//返回一个整数 

public int getTotalNumberOfEmployees() { 


return getJdbcTemplate().queryForInt("SELECT COUNT(0) FROM employees"); 

} 


//更新操作 

this.jdbcTemplate.update( 


"insert into t_actor (first_name, surname) values (?, ?)", 


new Object[] {"Leonor", "Watling"});
//返回值为一个长整形
public long getAverageAge() {
    return getJdbcTemplate().queryForLong("SELECT AVG(age) FROM employee");
  }
//返回一个整数
public int getTotalNumberOfEmployees() {
    return getJdbcTemplate().queryForInt("SELECT COUNT(0) FROM employees");
  }

//更新操作
this.jdbcTemplate.update(
        "insert into t_actor (first_name, surname) values (?, ?)", 
        new Object[] {"Leonor", "Watling"});

2.2、spring 2.5新功能,另类的jdbc ORM:BeanPropertyRowMapper

上面我们检索时必须实现RowMapper,将结果集转化为java对象。Spring2.5 简化了这一操作,使得我们不必再实现RowMapper,实现此功能的俩个神奇东东便是:ParameterizedRowMapper,ParameterizedBeanPropertyRowMapper,貌似通过java反射机制实现了将resultset字段映射到java对象,但是数据表的列必须和java对象的属性对应,没有研究源码,有点类似于apache 的BeanUtil,不知为何这部分在spring开发参考手册没有,难道不是经典。

//使用ParameterizedBeanPropertyRowMapper 

@SuppressWarnings({"unchecked"}) 


public List<Customer> getAll() { 


return getJdbcTemplate().query("select * from t_customer", ParameterizedBeanPropertyRowMapper.newInstance(Customer.class)); 

} 


//使用BeanPropertyRowMapper 

@SuppressWarnings({"unchecked"}) 


public List<Customer> getAll() { 


return getJdbcTemplate().query("select * from t_customer", new BeanPropertyRowMapper(Customer.class)); 

}
//使用ParameterizedBeanPropertyRowMapper
@SuppressWarnings({"unchecked"})
  public List<Customer> getAll() {
      return getJdbcTemplate().query("select * from t_customer", ParameterizedBeanPropertyRowMapper.newInstance(Customer.class));
  }

//使用BeanPropertyRowMapper
@SuppressWarnings({"unchecked"})
  public List<Customer> getAll() {
      return getJdbcTemplate().query("select * from t_customer", new BeanPropertyRowMapper(Customer.class));
  }

注意:ParameterizedBeanPropertyRowMapper是BeanPropertyRowMapper子类。另外表的字段名称必须和实体类的成员变量名称一致;

2.3、spring之JDBC批量操作

jdbcTemplate.batchUpdate(final String[] sql) ,API解释:Issue multiple SQL updates on a single JDBC Statement using batching,翻译过来大致为:解决多个sql的插入、更新、删除操作在一个Statement中。性能一般。

jdbcTemplate.batchUpdate(String sql, final BatchPreparedStatementSetter pss),类似于JDBC的PreparedStatement,性能较上着有所提高。

我们举例说明如何使用,示例如下:

final int count = 2000; 

final List<String> firstNames = new ArrayList<String>(count); 


final List<String> lastNames = new ArrayList<String>(count); 


for (int i = 0; i < count; i++) { 


firstNames.add("First Name " + i); 


lastNames.add("Last Name " + i); 

} 
jdbcTemplate.batchUpdate( 

"insert into customer (id, first_name, last_name, last_login, comments) values (?, ?, ?, ?, ?)", 


new BatchPreparedStatementSetter() { 

//为prepared statement设置参数。这个方法将在整个过程中被调用的次数 

public void setValues(PreparedStatement ps, int i) throws SQLException { 


ps.setLong(1, i + 10); 


ps.setString(2, firstNames.get(i)); 


ps.setString(3, lastNames.get(i)); 


ps.setNull(4, Types.TIMESTAMP); 


ps.setNull(5, Types.CLOB); 

} 
//返回更新的结果集条数 

public int getBatchSize() { 


return count; 

} 
}); 
}
final int count = 2000;
    final List<String> firstNames = new ArrayList<String>(count);
    final List<String> lastNames = new ArrayList<String>(count);
    for (int i = 0; i < count; i++) {
      firstNames.add("First Name " + i);
      lastNames.add("Last Name " + i);
    }
    jdbcTemplate.batchUpdate(
            "insert into customer (id, first_name, last_name, last_login, comments) values (?, ?, ?, ?, ?)",
            new BatchPreparedStatementSetter() {
           //为prepared statement设置参数。这个方法将在整个过程中被调用的次数
        public void setValues(PreparedStatement ps, int i) throws SQLException {
                ps.setLong(1, i + 10);
                ps.setString(2, firstNames.get(i));
                ps.setString(3, lastNames.get(i));
                ps.setNull(4, Types.TIMESTAMP);
                ps.setNull(5, Types.CLOB);
              }
              //返回更新的结果集条数
          public int getBatchSize() {
                   return count;
              }
            });
  }

BatchSqlUpdate类是SqlUpdate 的子类,适用于插入、删除、更新批量操作,内部使用PreparedStatement,所以效率很高,批量语句达到设定的batchSize,或者手动调用flush才会执行批量操作。注意:此类是非线程安全的,必须为每个使用者创建一个实例,或者在同一个线程中使用前调用reset。

下面我们举例说明如何使用BatchSqlUpdate,来执行批量操作。示例如下:

class BatchInsert extends BatchSqlUpdate { 

private static final String SQL = "insert into t_customer (id, first_name, last_name, last_login, " 


+ "comments) values (?, ?, ?, ?, null)"; 



BatchInsert(DataSource dataSource) { 

super(dataSource, SQL); 


declareParameter(new SqlParameter(Types.INTEGER)); 


declareParameter(new SqlParameter(Types.VARCHAR)); 


declareParameter(new SqlParameter(Types.VARCHAR)); 


declareParameter(new SqlParameter(Types.TIMESTAMP)); 




setBatchSize(10); 

} 


}
class BatchInsert extends BatchSqlUpdate {
  private static final String SQL = "insert into t_customer (id, first_name, last_name, last_login, "
      + "comments) values (?, ?, ?, ?, null)";

  BatchInsert(DataSource dataSource) {
    super(dataSource, SQL);
    declareParameter(new SqlParameter(Types.INTEGER));
    declareParameter(new SqlParameter(Types.VARCHAR));
    declareParameter(new SqlParameter(Types.VARCHAR));
    declareParameter(new SqlParameter(Types.TIMESTAMP));

    setBatchSize(10);
  }

}
int count = 5000; 

for (int i = 0; i < count; i++) { 


batchInsert.update(new Object[] { i + 100L, "a" + i, "b" + i, null }); 

}
int count = 5000;
    for (int i = 0; i < count; i++) {
      batchInsert.update(new Object[] { i + 100L, "a" + i, "b" + i, null });
    }

至此,spring JDBC主要的应用基本上都简单罗列一番,所有代码均为文章举例,不是很严谨,仅为演示每一种用法,抛砖引玉,希望有独特见解的拍砖,有问题的请指明问题所在,谢谢

相关推荐