为什么 ObjectOutputStream 和 ObjectInputStream 抛出 EOFException 的这种奇怪行为？

Question

我写了一个自定义 serializing/de-serializing 逻辑来保存一些数据，因为 Java 默认序列化结果是时间和内存都非常昂贵。为此，我为需要持久化的 class(es) 编写了 readObject(ObjectInput in) 和 writeObject(ObjectOutput out) 方法。但是我注意到如果我在 writeObject(ObjectOutput out) 方法中不使用任何 out.writeObject(obj) 那么它总是抛出 EOFException.

考虑以下示例：

Data.java

public class Data implements BaseData {

private String messageUID;
private String rawData;
private String data;
private Long type;
private Boolean processed = false;
private String processedMessage;
private String processedDetaildMessage;

// getter setter

public void readObject(ObjectInput in) throws IOException, ClassNotFoundException {
    messageUID = in.readUTF();
    rawData = in.readUTF();
    data = in.readUTF();
    type = in.readLong();
    processed = in.readBoolean();
    if (processed) {
        processedMessage = in.readUTF();
        processedDetaildMessage = in.readUTF();
    }
}

public void writeObject(ObjectOutput out) throws IOException {
    out.writeUTF(messageUID);
    out.writeUTF(rawData);
    out.writeUTF(data);
    out.writeLong(type);
    out.writeBoolean(processed);
    if (processed) {
        out.writeUTF(processedMessage);
        String tempDetailsMessage[] = processedDetaildMessage.split(" more");
        out.writeUTF(tempDetailsMessage[tempDetailsMessage.length - 1]);
    }
}

然而，每当我使用上面的代码时，out 流总是在末尾缺少一些信息（来自 processedDetaildMessage 字段），我在从 [=23] 读取它时得到 EOFException =]，下面的堆栈跟踪（Data.java 第 216 行是 processedDetaildMessage = in.readUTF()）；

java.io.EOFException
    at java.io.ObjectInputStream$BlockDataInputStream.readByte(ObjectInputStream.java:2766)
    at java.io.ObjectInputStream$BlockDataInputStream.readUTFChar(ObjectInputStream.java:3158)
    at java.io.ObjectInputStream$BlockDataInputStream.readUTFBody(ObjectInputStream.java:3055)
    at java.io.ObjectInputStream$BlockDataInputStream.readUTF(ObjectInputStream.java:2864)
    at java.io.ObjectInputStream.readUTF(ObjectInputStream.java:1072)
    at com.smartstream.common.Data.readObject(Data.java:216)
    at com.smartstream.common.PerformanceTest.getObjectFromBytes(PerformanceTest.java:168)
    at com.smartstream.common.PerformanceTest.access[=11=](PerformanceTest.java:160)
    at com.smartstream.common.PerformanceTest.mapRow(PerformanceTest.java:119)
    at com.smartstream.common.PerformanceTest.mapRow(PerformanceTest.java:1)
    at org.springframework.jdbc.core.RowMapperResultSetExtractor.extractData(RowMapperResultSetExtractor.java:92)
    at org.springframework.jdbc.core.RowMapperResultSetExtractor.extractData(RowMapperResultSetExtractor.java:60)
    at org.springframework.jdbc.core.JdbcTemplate.doInPreparedStatement(JdbcTemplate.java:651)
    at org.springframework.jdbc.core.JdbcTemplate.execute(JdbcTemplate.java:589)
    at org.springframework.jdbc.core.JdbcTemplate.query(JdbcTemplate.java:639)
    at org.springframework.jdbc.core.JdbcTemplate.query(JdbcTemplate.java:668)
    at org.springframework.jdbc.core.JdbcTemplate.query(JdbcTemplate.java:676)
    at org.springframework.jdbc.core.JdbcTemplate.queryForObject(JdbcTemplate.java:731)
    at com.smartstream.common.PerformanceTest.readFromDb(PerformanceTest.java:109)
    at com.smartstream.common.PerformanceTest.main(PerformanceTest.java:66)

所以我虽然会在写完所有必填字段后在末尾放置一些额外的 byte/s 信息，但不会读取它们，这样我在读取时就不会到达文件末尾。我尝试了所有这些 out.writeByte(-1)、out.writeInt(-1)、out.writeLong(2342343l)、out.writeUTF("END_OF_STREAM")，但这些都没有区别。最后我做了 out.writeObject(new String("END_OF_STREAM")) 并且效果很好。如果 none 的信息是使用 writeObject() 方法编写的，有人可以解释为什么 outputstream 会遗漏一些信息吗？下面是我如何读写 to/from streams;

private byte[] getObjectAsBytes(Data data) {
    byte[] byteArray = null;
    ByteArrayOutputStream bos = null;
    ObjectOutputStream oos = null;
    try {
        bos = new ByteArrayOutputStream();
        oos = new ObjectOutputStream(bos);
        // Use this for java default serialization
        // oos.writeObject(data);
        data.writeObject(oos);
        byteArray = bos.toByteArray();
    } catch (IOException e) {
        e.printStackTrace();
    } finally {
        if (oos != null) {
            try {
                oos.flush();
                oos.close();
            } catch (IOException e) {
                e.printStackTrace();
            }
        }
    }
    return byteArray;
}

private Data getObjectFromBytes(byte[] byteArray) {
    Data data = new Data();
    ByteArrayInputStream bais = null;
    ObjectInputStream ois = null;
    try {
        bais = new ByteArrayInputStream(byteArray);
        ois = new ObjectInputStream(bais);
        // Use this for java default serialization
        // data = (Data) ois.readObject();
        data.readObject(ois);
    } catch (IOException e) {
        e.printStackTrace();
    } catch (ClassNotFoundException e) {
        e.printStackTrace();
    } finally {
        if (ois != null) {
            try {
                ois.close();
            } catch (IOException e) {
                e.printStackTrace();
            }
        }
    }
    return data;
}

如果有人感兴趣，下面是流中写的内容；

使用原始代码保留数据（抛出 EOFException 和缺失信息）（不要将堆栈跟踪与原始问题混淆，此堆栈跟踪作为字段保留 processedDetailedMessage)

¬í---z-------3507319347632941385----FEEDER-----1437052314954 ---This is a random string---N---þ%J---!this is message of processed dataÛ
Caused by: java.sql.SQLException: ORA-01691: unable to extend lob segment TLM_DBO.SYS_LOB0000076335C00008$$ by 8192 in tablespace WIN_SL_TABLE
at oracle.jdbc.driver.T4CTTIoer.processError(T4CTTIoer.java:439)
at oracle.jdbc.driver.T4CTTIoer.processError(T4CTTIoer.java:395)
at oracle.jdbc.driver.T4C8Oall.processError(T4C8Oall.java:802)
at oracle.jdbc.driver.T4CTTIfun.receive(T4CTTIfun.java:436)
at oracle.jdbc.driver.T4CTTIfun.doRPC(T4CTTIfun.java:186)
at oracle.jdbc.driver.T4C8Oall.doOALL(T4C8Oall.java:521)
at oracle.jdbc.driver.T4CPreparedStatement.doOall8(T4CPreparedStatement.java:205)
at oracle.jdbc.driver.T4CPreparedStatement.executeForRows(T4CPreparedStatement.java:1008)
at oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java:1307)
at oracle.jdbc.driver.OraclePreparedStatement.executeInternal(OraclePreparedStatement.java:3449)
at oracle.jdbc.driver.OraclePre

使用writeObject方法在末尾写入额外字符串后保留数据

¬í---z-------3507319347632941385----FEEDER-----1437052314954 ---This is a random string---N---þ%J---!this is message of processed dataÛ
Caused by: java.sql.SQLException: ORA-01691: unable to extend lob segment TLM_DBO.SYS_LOB0000076335C00008$$ by 8192 in tablespace WIN_SL_TABLE
at oracle.jdbc.driver.T4CTTIoer.processError(T4CTTIoer.java:439)
at oracle.jdbc.driver.T4CTTIoer.processError(T4CTTIoer.java:395)
at oracle.jdbc.driver.T4C8Oall.processError(T4C8Oall.java:802)
at oracle.jdbc.driver.T4CTTIfun.receive(T4CTTIfun.java:436)
at oracle.jdbc.driver.T4CTTIfun.doRPC(T4CTTIfun.java:186)
at oracle.jdbc.driver.T4C8Oall.doOALL(T4C8Oall.java:521)
at oracle.jdbc.driver.T4CPreparedStatement.doOall8(T4CPreparedStatement.java:205)
at oracle.jdbc.driver.T4CPreparedStatement.executeForRows(T4CPreparedStatement.java:1008)
at oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java:1307)
at oracle.jdbc.driver.OraclePreparedStatement.executeInternal(OraclePreparedStatement.java:3449)
at oracle.jdbc.driver.OraclePrz-----NeparedStatement.execute(OraclePreparedStatement.java:3550)
at oracle.jdbc.driver.OraclePreparedStatementWrapper.execute(OraclePreparedStatementWrapper.java:1374)
at com.ibm.ws.rsadapter.jdbc.WSJdbcPreparedStatement.pmiExecute(WSJdbcPreparedStatement.java:975)
at com.ibm.ws.rsadapter.jdbc.WSJdbcPreparedStatement.execute(WSJdbcPreparedStatement.java:642)
at com.smartstream.control.engine.config.dao.jdbc.ProcessExecutionAuditDetailDao.doInPreparedStatement(ProcessExecutionAuditDetailDao.java:115)
at org.springframework.jdbc.core.JdbcTemplate.execute(JdbcTemplate.java:586)
... 23t 
END_OF_STREAM

PS ----表示不可读字节

Answer 1

您的持久数据不完整，因为您在刷新 ObjectOutputStream 之前创建字节数组。在 getObjectAsBytes(Data) 中，在 finally 块之后移动 byteArray = bos.toByteArray(); 以使其工作。或者，该方法可以更简洁地编写如下（需要 Java 7+）：

private byte[] getObjectAsBytes(Data data) { ByteArrayOutputStream bos = new ByteArrayOutputStream(); try (ObjectOutputStream oos = new ObjectOutputStream(bos)) { data.writeObject(oos); } catch (IOException e) { e.printStackTrace(); } return bos.toByteArray(); }

我在自己的程序中测试了这两种方式，它们都可以防止抛出 EOFException。

至于 writeObject 为何有效，那是因为 underlying writeObject implementation toggles block data mode at the beginning and ending of the method, and changing the block data mode performs a drain 将所有数据写入底层 OutputStream，对于 ByteArrayOutputStream，它实际上与刷新相同。

Answer 2

此问题是由于 writeObject 方法和其他一些 non-generic write* 方法（即 writeUTF）的不同实现引起的。 writeObject 方法在方法开始和结束时切换到数据块模式，这导致所有数据都写入底层 OutputStream，这与在 outputStream 上调用 flush 具有相同的效果。这意味着您不能在将剩余数据刷新到流之前创建另一个 byteArray。最好暂时坚持使用 writeObject 方法；即

public void writeObject(ObjectOutput out) throws IOException {
    out.writeUTF(messageUID);
    out.writeUTF(rawData);
    out.writeUTF(data);
    out.writeLong(type);
    out.writeBoolean(processed);
    if (processed) {
        out.writeUTF(processedMessage);
        String tempDetailsMessage[] = processedDetaildMessage.split(" more");
        out.writeObject(tempDetailsMessage[tempDetailsMessage.length - 1]);
    }
}

为什么 ObjectOutputStream 和 ObjectInputStream 抛出 EOFException 的这种奇怪行为？

Why this strange behaviour of ObjectOutputStream and ObjectInputStream throwing EOFException?

java

objectoutputstream

objectinputstream