CP502: Non-Text Files.

4.2. Non-Text Files.

The previous example also showed the use of binary copying of a file - it did not assume that the file contained line by line characters. It repetitively read a large number of bytes and wrote those bytes to the second file.

It is possible to have a file of any data type.

General files are binary files: i.e. data is not stored in human readable form.

  FileOutputStream out_stream= new FileOutputStream("fred.dat");    // Get a FileOutputStream object


  FileInputStream in_stream = new FileInputStream("freda.dat");    // Get a FileInputStream object

This class will let you read in / write out byte information. Perhaps better and more often used is:

  FileOutputStream out_stream= new FileOutputStream("fred.dat");    // Get a FileOutputStream object

  DataOutputStream data_out_stream= new DataOutputStream(out_stream);    // Get a DataOutputStream object


  FileInputStream in_stream = new FileInputStream("freda.dat");    // Get a FileInputStream object

  DataInputStream data_in_stream = new DataInputStream(in_stream);      // Get a DataInputStream object

These will let you read and write simple types as binary data via pre-defined methods:

 
                         DataInputStream


  read(byte[])

    Reads data into an array of bytes.

  read(byte[], int, int)

    Reads data into an array of bytes.

  readBoolean()

    Reads a boolean.

  readByte()

    Reads an 8 bit byte.

  readChar()

    Reads a 16 bit char.

  readDouble()

    Reads a 64 bit double.

  readFloat()

    Reads a 32 bit float.

  readFully(byte[])

    Reads bytes, blocking until all bytes are read.

  readFully(byte[], int, int)

    Reads bytes, blocking until all bytes are read.

  readInt()

    Reads a 32 bit int.

  readLine()

    Reads in a line that has been

      terminated by a \n, \r, \r\n or EOF.

  readLong()

    Reads a 64 bit long.

  readShort()

    Reads a 16 bit short.

  readUTF()

    Reads a UTF format String.

  readUTF(DataInput)

    Reads a UTF format String from

       the given input stream.

  readUnsignedByte()

    Reads an unsigned 8 bit byte.

  readUnsignedShort()

    Reads 16 bit short.

  skipBytes(int)

    Skips bytes, blocks until all bytes are skipped.

 
                        DataOutputStream


  flush()

    Flushes the stream.

  size()

    Returns the number of bytes written.

  write(int)

    Writes a byte.

  write(byte[], int, int)

    Writes a sub array of bytes.

  writeBoolean(boolean)

    Writes a boolean.

  writeByte(int)

    Writes an 8 bit byte.

  writeBytes(String)

    Writes a String as a sequence of bytes.

  writeChar(int)

    Writes a 16 bit char.

  writeChars(String)

    Writes a String as a sequence of chars.

  writeDouble(double)

    Writes a 64 bit double.

  writeFloat(float)

    Writes a 32 bit float.

  writeInt(int)

    Writes a 32 bit int.

  writeLong(long)

    Writes a 64 bit long.

  writeShort(int)

    Writes a 16 bit short.

  writeUTF(String)

    Writes a String in UTF format.

A non - text file contains binary data (ie machine readable). This means that each file record will use up as many bytes as is required for that data type.

4.2.1. Random Access

Java has the facility to use Random Access files - a record may be read from or written to any place in the file without having had to access the preceding records. Typically, a seek routine is used to position the file pointer at the right place for a subsequent read or write.

The RandomAccessFile class has the following constructors:

public RandomAccessFile(String name,

                          String mode) throws IOException


      Creates a RandomAccessFile with the specified system dependent file name and the

      specified mode. Mode "r" is for read-only and mode "rw" is for read+write.

      Parameters:

            name - the system dependent file name

            mode - the access mode

      Throws: IOException

            If an I/O error has occurred.


  public RandomAccessFile(File file,

                          String mode) throws IOException


      Creates a RandomAccessFile from a specified File object and mode ("r" or "rw").

      Parameters:

            file - the file object

            mode - the access mode

A RandomAccessFile supports the previous binary read/write capabilities as well as the following methods:

  public long getFilePointer() throws IOException


      Returns the current location of the file pointer.


  public void seek(long pos) throws IOException


      Sets the file pointer to the specified absolute position.

      Parameters:

            pos - the absolute position

This enables a user to get or set the position within the RandomAccessFile.

4.2.2. Writing records in Java.

How would we store records in a file using Java? We would like to be able to write out the entire record into a file and also to be able to read the entire record back. We could use:

import local.units.sdsu.io.*;

import java.lang.*;

import java.io.*;


class Record

{

     public int studentNumber;

     public String firstName;

     public String lastName;

     public double examMark;

}


public class RecordTest

{

     public static void main(String argv[]) throws IOException

     {

     Record fred = new Record();

     RandomAccessFile file = new RandomAccessFile("freda.dat","rw");

     byte firstName[] ;

     byte lastName[] ;


          fred.studentNumber = 730372;

          fred.firstName = "Andrew";

          fred.lastName  = "Marriott";

          fred.examMark  = 27.3;


          file.writeInt(fred.studentNumber);


          firstName = new byte[fred.firstName.length()];

          fred.firstName.getBytes( 0, fred.firstName.length(),firstName,0);

          file.write(firstName);


          lastName = new byte[fred.lastName.length()];

          fred.lastName.getBytes( 0, fred.lastName.length(),lastName,0);

          file.write(lastName);

          file.writeDouble(fred.examMark);

          file.close();

        }

}

But notice that this is not the normal way to do this in Java - we use methods within the class to read and write the classes data. Note as well that the size of each record is now variable because the String types have a variable number of characters. This is undesirable in a RandomAccessFile as we want to have records of exactly the same size. We do not know what the size is!!! We could restrict the size of each string printed out to some arbitrary value:

          firstName = new byte[80];

          fred.firstName.getBytes( 0, fred.firstName.length(),firstName,0);

          file.write(firstName);

          // Write out 80 bytes each time - what is the padding?

Now our record is always a fixed size. When we read the record we know we should :

          fred.studentNumber = file.readInt();


          byte firstname[] = new byte[80];

          file.readFully( firstname );

          fred.firstName = new String( firstname, 0 );


          byte lastname[] = new byte[80];

          file.readFully( lastname );

          fred.lastName = new String( lastname, 0 );


          fred.examMark = file.readDouble();

Instead of using a value we would use a const with an appropriate name. And to calculate the size :

     public int size() // Size of a Record.

     {

          return (

               4 +       // size of an integer in bytes

               stringSize +   // size of firstname

               stringSize +   // size of lastname

               8         // size of a double in bytes.

               );

     }

Although this hard coding of a calculated value is not good practice. The entire class for reading a record looks like:

import local.units.sdsu.io.*;

import java.lang.*;

import java.io.*;


public class Client

{

// Mainline Variable Declarations

static PrintStream output = System.out;

static ASCIIInputStream input = new ASCIIInputStream(System.in);


protected int studentNumber;

protected String firstName;

protected String lastName;

protected double examMark;


private static final int stringSize = 80;


private static RandomAccessFile raf = null;


private static final String NO_FILENAME_SET = "No Filename Set!";


     public static void openClientFile(String name)

          throws IOException

     {

          raf = new RandomAccessFile(name,"r");

     }


     public static void openClientFile(File file)

          throws IOException

     {

          raf = new RandomAccessFile(file,"r");

     }


     public static void createClientFile(File file)

          throws IOException

     {

          raf = new RandomAccessFile(file,"rw");

     }


     public static void createClientFile(String name)

          throws IOException

     {

          raf = new RandomAccessFile(name,"rw");

     }


     public static void close()

          throws IOException

     {

          if(raf != null)

               raf.close();

          raf = null;

     }


     public static long getFilePointer()

          throws IOException

     {

          if(raf != null)

               return raf.getFilePointer();

          throw  new IOException(NO_FILENAME_SET);

     }

     public static long length()

          throws IOException

     {

          if(raf != null)

               return raf.length();

          throw  new IOException(NO_FILENAME_SET);

     }


     public static void seek(long pos)

          throws IOException

     {

          if(raf != null)

               raf.seek(pos);

          throw  new IOException(NO_FILENAME_SET);

     }


     public String toString()

     {

          return(studentNumber + ":" + firstName +

               " " + lastName + ":" + examMark + ":");

     }

    public Client(  int studentnumber,

          String firstname,

          String lastname,

          double exammark)

    {

        studentNumber = studentnumber;

        firstName = firstname;

        lastName = lastname;

        examMark = exammark;

    }


    // Read a Client record from the specified RandomAccessFile

    public void getClient() throws IOException

    {

        if(raf == null)

          throw new IOException(NO_FILENAME_SET);


        studentNumber = raf.readInt();


    byte  fn[] = new byte[stringSize];

        raf.readFully(fn);

        firstName = new String(fn,0);


    byte  ln[] = new byte[stringSize];

        raf.readFully(ln);

        lastName = new String(ln,0);


        examMark = raf.readDouble();

    }


    // Write a record to the specified RandomAccessFile

    public void putClient() throws IOException

    {

        if(raf == null)

          throw new IOException(NO_FILENAME_SET);


        raf.writeInt( studentNumber );


    byte  fn[] = new byte[stringSize];


        if (firstName != null )

          firstName.getBytes(0,firstName.length(),fn,0);


        raf.write(fn);


    byte  ln[] = new byte[stringSize];


        if ( lastName != null )

          lastName.getBytes(0,lastName.length(),ln,0);


        raf.write(ln);


        raf.writeDouble( examMark );

    }


     // Size of a Record.

     public int size()

    {

        return (

          4 +       // size of an integer in bytes

          stringSize +   // size of firstname

          stringSize +   // size of lastname

          8         // size of a double in bytes.

          );

    }


    public static void main(String argv[]) throws IOException

    {

    Client client = new Client(730372,"Andrew","Marriott",27.3);


        client.createClientFile("fred3.dat");


        client.putClient();


        Format.print(output, client + "0);

        client.close();

    }

}

It is possible for you to grab the source code and compile it on your machine via

Client.java

After retrieving and saving the code, you should be able to type:
javac Client.java ; java Client
This will execute the code. People not at Curtin will need to setup their Java environment first.

Note that many of the methods simply call the appropriate RandomAccessFile methods.

Another class which calculates the record size dynamically and reads/writes the data in that size chunks can be designed and the differences between our first attempt and this one are shown below. This class uses the ByteArrayInputStream and ByteArrayOutputStream classes.

public class Client2

{

          :

private static final int stringSize = 80;

private static int recordSize = -1;

private static DataInputStream din = null;

private static ByteArrayInputStream bais = null;

private static DataOutputStream dout = null;

private static ByteArrayOutputStream baos = null;


          :


    // Read a Client record from the specified RandomAccessFile

    public void getClient() throws IOException

    {

     if(raf == null)

          throw new IOException(NO_FILENAME_SET);


    byte  record[] = new byte[size()];

     raf.readFully(record,0,recordSize);


     bais = new ByteArrayInputStream(record);

     din = new DataInputStream(bais);


     studentNumber = din.readInt();


    byte  fn[] = new byte[stringSize];

     din.readFully(fn);

     firstName = new String(fn,0);


    byte  ln[] = new byte[stringSize];

     din.readFully(ln);

     lastName = new String(ln,0);


     examMark = din.readDouble();

    }


    // Write a record to the specified RandomAccessFile

    public void putClient() throws IOException

    {

     putClientIntoBAOS();

     raf.write(baos.toByteArray());

     baos.reset();

    }

    public void putClientIntoBAOS()

     throws IOException

    {

     if(raf == null)

          throw new IOException(NO_FILENAME_SET);


     if ( baos == null)

     {

          baos = new ByteArrayOutputStream();

          dout = new DataOutputStream(baos);

     }


     dout.writeInt( studentNumber );


    byte  fn[] = new byte[stringSize];


     if (firstName != null )

          firstName.getBytes(0,firstName.length(),fn,0);


     dout.write(fn);


    byte  ln[] = new byte[stringSize];


     if ( lastName != null )

          lastName.getBytes(0,lastName.length(),ln,0);


     dout.write(ln);


     dout.writeDouble( examMark );


     recordSize = baos.size();

    }


     // Size of a Record.

     public int size()

    {

     if(recordSize == -1)

     {

          try

               putClientIntoBAOS();

          catch (IOException e)

          {

               recordSize = 0;

          }

          baos.reset();

     }

     return (recordSize);

    }

     :

}

It is possible for you to grab the source code and compile it on your machine via

Client2.java

After retrieving and saving the code, you should be able to type:
javac Client2.java ; java Client2
This will execute the code. People not at Curtin will need to setup their Java environment first.

So we can use files for storing large (or small) temporary results just like an array (although in some languages, we can only access it sequentially). This is often useful when we have a large piece of text or data that would not fit into memory at once, that we need to process a few times. e.g. text editor, compiler, assembler, etc.

 import local.units.sdsu.io.*;

 import java.lang.*;

 import java.io.*;


 class Client;

 {

 // Data of the client record.


 // Methods to read and write the client data.

 }


 public class MatchMaker

 {

 // Mainline Variable Declarations

 static PrintStream output = System.out;

 static ASCIIInputStream input = new ASCIIInputStream(System.in);


 static Client match, applicant  ;

 static int count;


     public static void main(String argv[]) throws IOException

     {

                   : // Get information about applicant

       count = 0;

       match.openClientFile("clientel.dat");

       applicant.createClientFile("possibles.dat");

       while ( !match.eof() )

       {

             /* get a client from file to match against applicant */

             match.getClient();

             if( (applicant.sex !=  match.sex) &&

                   (applicant.age - match.age<=5) &&

                   (                         )

                   (                )

             {

                   count++;

                   applicant.putClient();

             }

       }


       if (count==0) Format.print("Better became a nun or monk");

       else

       {

             applicant.close();

             applicant.openClientFile("possibles.dat");

             while ( !applicant.eof() )

             {

                   applicant.getClient();

                   /* printout vital information */

             }

       }

     }

 }