JAVA完全控制Oracle中BLOB、CLOB说明

转载：
http://blog.csdn.net/wazj0517/archive/2006/10/19/1341074.aspx


环境：
Database: Oracle 9i
App Server: BEA Weblogic 8.14
表结构：
CREATE TABLE TESTBLOB (ID Int, NAME Varchar2(20), BLOBATTR Blob)
CREATE TABLE TESTBLOB (ID Int, NAME Varchar2(20), CLOBATTR Clob)

JAVA可以通过JDBC，也可以通过JNDI访问并操作数据库，这两种方式的具体操作存在着一些差异，由于通过App Server的数据库连接池JNDI获得的数据库连接提供的java.sql.Blob和java.sql.Clob实现类与JDBC方式提供的不同，因此在入库操作的时候需要分别对待；出库操作没有这种差异，因此不用单独对待。

一、BLOB操作
1、入库
（1）JDBC方式
    //通过JDBC获得数据库连接
    Class.forName("oracle.jdbc.driver.OracleDriver");
    Connection con = DriverManager.getConnection("jdbc:oracle:thin:@localhost:1521:testdb", "test", "test");
    con.setAutoCommit(false);
    Statement st = con.createStatement();
    //插入一个空对象empty_blob()
    st.executeUpdate("insert into TESTBLOB (ID, NAME, BLOBATTR) values (1, "thename", empty_blob())");
    //锁定数据行进行更新，注意“for update”语句
    ResultSet rs = st.executeQuery("select BLOBATTR from TESTBLOB where ID=1 for update");
    if (rs.next())
    {
        //得到java.sql.Blob对象后强制转换为oracle.sql.BLOB
        oracle.sql.BLOB blob = (oracle.sql.BLOB) rs.getBlob("BLOBATTR");
        OutputStream outStream = blob.getBinaryOutputStream();
        //data是传入的byte数组，定义：byte[] data
        outStream.write(data, 0, data.length);
    }
    outStream.flush();
    outStream.close();
    con.commit();
    con.close();
（2）JNDI方式
    //通过JNDI获得数据库连接
    Context context = new InitialContext();
    ds = (DataSource) context.lookup("ORA_JNDI");
    Connection con = ds.getConnection();
    con.setAutoCommit(false);
    Statement st = con.createStatement();
    //插入一个空对象empty_blob()
    st.executeUpdate("insert into TESTBLOB (ID, NAME, BLOBATTR) values (1, "thename", empty_blob())");
    //锁定数据行进行更新，注意“for update”语句
    ResultSet rs = st.executeQuery("select BLOBATTR from TESTBLOB where ID=1 for update");
    if (rs.next())
    {
        //得到java.sql.Blob对象后强制转换为weblogic.jdbc.vendor.oracle.OracleThinBlob（不同的App Server对应的可能会不同）
        weblogic.jdbc.vendor.oracle.OracleThinBlob blob = (weblogic.jdbc.vendor.oracle.OracleThinBlob) rs.getBlob("BLOBATTR");
        OutputStream outStream = blob.getBinaryOutputStream();
        //data是传入的byte数组，定义：byte[] data
        outStream.write(data, 0, data.length);
    }
    outStream.flush();
    outStream.close();
    con.commit();
    con.close();
2、出库
    //获得数据库连接
    Connection con = ConnectionFactory.getConnection();
    con.setAutoCommit(false);
    Statement st = con.createStatement();
    //不需要“for update”
    ResultSet rs = st.executeQuery("select BLOBATTR from TESTBLOB where ID=1");
    if (rs.next())
    {
        java.sql.Blob blob = rs.getBlob("BLOBATTR");
        InputStream inStream = blob.getBinaryStream();
        //data是读出并需要返回的数据，类型是byte[]
        data = new byte[input.available()];
        inStream.read(data);
        inStream.close();
    }
    inStream.close();
    con.commit();
    con.close();

二、CLOB操作
1、入库
（1）JDBC方式
    //通过JDBC获得数据库连接
    Class.forName("oracle.jdbc.driver.OracleDriver");
    Connection con = DriverManager.getConnection("jdbc:oracle:thin:@localhost:1521:testdb", "test", "test");
    con.setAutoCommit(false);
    Statement st = con.createStatement();
    //插入一个空对象empty_clob()
    st.executeUpdate("insert into TESTCLOB (ID, NAME, CLOBATTR) values (1, "thename", empty_clob())");
    //锁定数据行进行更新，注意“for update”语句
    ResultSet rs = st.executeQuery("select CLOBATTR from TESTCLOB where ID=1 for update");
    if (rs.next())
    {
        //得到java.sql.Clob对象后强制转换为oracle.sql.CLOB
        oracle.sql.CLOB clob = (oracle.sql.CLOB) rs.getClob("CLOBATTR");
        Writer outStream = clob.getCharacterOutputStream();
        //data是传入的字符串，定义：String data
        char[] c = data.toCharArray();
        outStream.write(c, 0, c.length);
    }
    outStream.flush();
    outStream.close();
    con.commit();
    con.close();
（2）JNDI方式
    //通过JNDI获得数据库连接
    Context context = new InitialContext();
    ds = (DataSource) context.lookup("ORA_JNDI");
    Connection con = ds.getConnection();
    con.setAutoCommit(false);
    Statement st = con.createStatement();
    //插入一个空对象empty_clob()
    st.executeUpdate("insert into TESTCLOB (ID, NAME, CLOBATTR) values (1, "thename", empty_clob())");
    //锁定数据行进行更新，注意“for update”语句
    ResultSet rs = st.executeQuery("select CLOBATTR from TESTCLOB where ID=1 for update");
    if (rs.next())
    {
        //得到java.sql.Clob对象后强制转换为weblogic.jdbc.vendor.oracle.OracleThinClob（不同的App Server对应的可能会不同）
        weblogic.jdbc.vendor.oracle.OracleThinClob clob = (weblogic.jdbc.vendor.oracle.OracleThinClob) rs.getClob("CLOBATTR");
        Writer outStream = clob.getCharacterOutputStream();
        //data是传入的字符串，定义：String data
        char[] c = data.toCharArray();
        outStream.write(c, 0, c.length);
    }
    outStream.flush();
    outStream.close();
    con.commit();
    con.close();
2、出库
    //获得数据库连接
    Connection con = ConnectionFactory.getConnection();
    con.setAutoCommit(false);
    Statement st = con.createStatement();
    //不需要“for update”
    ResultSet rs = st.executeQuery("select CLOBATTR from TESTCLOB where ID=1");
    if (rs.next())
    {
        java.sql.Clob clob = rs.getClob("CLOBATTR");
        Reader inStream = clob.getCharacterStream();
        char[] c = new char[(int) clob.length()];
        inStream.read(c);
        //data是读出并需要返回的数据，类型是String
        data = new String(c);
        inStream.close();
    }
    inStream.close();
    con.commit();
    con.close();

需要注意的地方：
1、java.sql.Blob、oracle.sql.BLOB、weblogic.jdbc.vendor.oracle.OracleThinBlob几种类型的区别
2、java.sql.Clob、oracle.sql.CLOB、weblogic.jdbc.vendor.oracle.OracleThinClob几种类型的区别
公司项目中的用法(博客)：
入库：先插一个oracle.sql.CLOB.empty_lob()进去，然后
String updateBaseSourceSql = "select content from mb_baseSource where id = ? for update";
    conn.setAutoCommit(false);
    ps = conn.prepareStatement(updateBaseSourceSql);
    ps.setLong(1, result);
    ResultSet rs = ps.executeQuery();
    oracle.sql.CLOB clob = null;
    if (rs.next()) {
     clob = (oracle.sql.CLOB) rs.getClob(1);
    }
    Writer wr = clob.getCharacterOutputStream();
    wr.write(baseSource[4]);
    wr.flush();
    wr.close();
    rs.close();
    ps.close();
    conn.commit();
出库：
findBaseSourceSql = "select content from mb_baseSource where id = ?";
   ps = conn.prepareStatement(findBaseSourceSql);
   ps.setLong(1, sourceID);
   rs = ps.executeQuery();
   if (rs.next()) {
    CLOB clob = (oracle.sql.CLOB) rs.getClob(1);
    if (clob != null) {
     Reader is = clob.getCharacterStream();
     BufferedReader br = new BufferedReader(is);
     String s = br.readLine();
     while (s != null) {
      result[6] += s;
      s = br.readLine();
     }
    }
   }
   rs.close();
   ps.close();
   conn.close();