Oracle 存储过程

简要记录存储过程语法与Java程序的调用方式

一 存储过程

首先，我们建立一个简单的表进行存储过程的测试

create table 
xuesheng(id integer, xing_ming varchar2(25), yu_wen number, shu_xue number);

insert into xuesheng values(1,'zhangsan',80,90)
insert into xuesheng values(2,'lisi',85,87)

1）无返回值的存储过程

create or replace procedure xs_proc_no is
begin
  insert into xuesheng values (3, 'wangwu', 90, 90);
  commit;
end xs_proc_no;

2）有单个数据值返回的存储过程

create or replace procedure xs_proc(temp_name in varchar2,
                                    temp_num  out number) is
  num_1 number;
  num_2 number;
begin
  select yu_wen, shu_xue
    into num_1, num_2
    from xuesheng
   where xing_ming = temp_name;
  --dbms_output.put_line(num_1 + num_2);
  temp_num := num_1 + num_2;
end;

其中，以上两种与sql server基本类似，而对于返回数据集时，上述方法则不能满足我们的要求。在Oracle中，一般使用ref cursor来返回数据集。示例代码如下：

3）有返回值的存储过程(列表返回)

首先，建立我们自己的包。并定义包中的一个自定义ref cursor

create or replace package mypackage as
  type my_cursor is ref cursor;
end mypackage;

在定义了ref cursor后，可以书写我们的程序代码

create or replace procedure xs_proc_list(shuxue   in number,
                                         p_cursor out mypackage.my_cursor) is
begin
  open p_cursor for
    select * from xuesheng where shu_xue > shuxue;
end xs_proc_list;

二、程序调用

在本节中，我们使用java语言调用存储过程。其中，关键是使用CallableStatement这个对象，代码如下：

String oracleDriverName = "oracle.jdbc.driver.OracleDriver";
 
        // 以下使用的Test就是Oracle里的表空间
        String oracleUrlToConnect = "jdbc:oracle:thin:@127.0.0.1:1521:orcl";
        Connection myConnection = null;
        try {
            Class.forName(oracleDriverName);
        } catch (ClassNotFoundException ex) {
            ex.printStackTrace();
        }
        try {
            myConnection = DriverManager.getConnection(oracleUrlToConnect,
                    "xxxx", "xxxx");//此处为数据库用户名与密码
 
        } catch (Exception ex) {
            ex.printStackTrace();
        }
        try {
             
            CallableStatement proc=null;
            proc=myConnection.prepareCall("{call xs_proc(?,?)}");
            proc.setString(1, "zhangsan");
            proc.registerOutParameter(2, Types.NUMERIC);
            proc.execute();
            String teststring=proc.getString(2);
            System.out.println(teststring);
 
        } catch (Exception ex) {
            ex.printStackTrace();
        }

对于列表返回值的存储过程，在上述代码中做简单修改。如下

CallableStatement proc=null;
            proc=myConnection.prepareCall("{call getdcsj(?,?,?,?,?)}");
            proc.setString(1, strDate);
            proc.setString(2, jzbh);
            proc.registerOutParameter(3, Types.NUMERIC);
            proc.registerOutParameter(4, OracleTypes.CURSOR);
            proc.registerOutParameter(5, OracleTypes.CURSOR);
            proc.execute();
            ResultSet rs=null;
            int total_number=proc.getInt(3);
            rs=(ResultSet)proc.getObject(4);

上述存储过程修改完毕。另外，一个复杂的工程项目中的例子：查询一段数据中间隔不超过十分钟且连续超过100条的数据。即上述代码所调用的getdcsj存储过程

create or replace procedure getDcsj(var_flag     in varchar2,
                                    var_jzbh     in varchar2,
                                    number_total out number,
                                    var_cursor_a out mypackage.my_cursor,
                                    var_cursor_b out mypackage.my_cursor) is
  total number;
  cursor cur is
    select sj, flag
      from d_dcsj
     where jzbh = var_jzbh
     order by sj desc
       for update;
  last_time date;
begin
  for cur1 in cur loop
    if last_time is null or cur1.sj >= last_time - 10 / 60 / 24 then
      update d_dcsj set flag = var_flag where current of cur;
      last_time := cur1.sj;
    else
      select count(*) into total from d_dcsj where flag = var_flag;
      dbms_output.put_line(total);
      if total < 100 then
        update d_dcsj set flag = null where flag = var_flag;
        last_time := null;
        update d_dcsj set flag = var_flag where current of cur;
      else
        open var_cursor_a for
          select *
            from d_dcsj
           where flag = var_flag
             and jzbh = var_jzbh
             and zh = 'A'
           order by sj desc;
        number_total := total;
        open var_cursor_b for
          select *
            from d_dcsj
           where flag = var_flag
             and jzbh = var_jzbh
             and zh = 'B'
           order by sj desc;
        number_total := total;
        exit;
      end if;
    end if;
  end loop;
  select count(*) into total from d_dcsj where flag = var_flag;
  dbms_output.put_line(total);
  if total < 100 then
    open var_cursor_a for
      select * from d_dcsj where zh = 'C';
    open var_cursor_b for
      select * from d_dcsj where zh = 'C';
  else
    open var_cursor_a for
      select *
        from d_dcsj
       where flag = var_flag
         and jzbh = var_jzbh
         and zh = 'A'
       order by sj desc;
    number_total := total;
    open var_cursor_b for
      select *
        from d_dcsj
       where flag = var_flag
         and jzbh = var_jzbh
         and zh = 'B'
       order by sj desc;
    number_total := total;
  end if;
  commit;
end;

本文转自：http://www.cnblogs.com/liliu/archive/2011/06/22/2087546.html