New Page 1

Reading from files and URLs

Problem:

There are lots of I/O classes

Best solution:

You carefully pick the right one

Not-as-good solution:

Someone picks the first class that catches their attention

Characters vs. Bytes

Classes for reading are in java.io.*

Some reading classes have "...Reader.." in the name

FileReader - for reading text files

BufferedReader - for "buffered" reading of text

- has a readLine() method

InputStreamReader - reads bytes, outputs chars

Etc.

If it has "...Reader..." in the name -

. . . it relates to text (bunches of chars)

Some reading classes have "...Stream..." in the name

InputStreamReader - (this has "Stream" AND "Reader" in it)

FileInputStream - reads bytes from a file

BufferedInputStream - buffers, so fewer disk accesses

DataInputStream - reads bytes, converts to ints, etc.

- has read(), readInt(), readBoolean(), etc., methods

Etc.

If it has "...Stream..." in the name -

It relates to bytes

Reading: Fast and Slow

The Theory:

Disk accesses are slow

To go faster, read(write) lots per disk access

To go slower, read(write) one byte/char/etc. per disk access

Solution 1: YOU read lots per disk access

int amount;
InputStream f = new
   FileInputStream("proteinNames1.txt");
amount=f.available();
byte buffer[]= new byte[amount];
if(f.read(buffer) != amount)
   System.err.println
     ("Couldn't read "+ amount + " bytes");

To go even slower, use a loop & read one byte per loop

Computer does a special disk access for each byte

Solution 2: tell the JAVA system to read lots per disk access

BufferedReader - "buffers" a text file

BufferedInputStream - "buffers" a data (byte) file

Buffering means:

The computer reads files in largish chunks

... stores the chunks in memory

You can read one byte/char at a time

...but you are getting them from the buffer

When buffer runs dry, computer reads another chunk

It's automatic - you don't have to worry about it