Transcript More on a First Java Program COMP 102 #4 17 July 2006

File handling and Scanning
COMP 112 2017T1
David Streader
Computer Science
Victoria University of Wellington
Copyright: David Streader, Victoria University of Wellington
Overview
COMP 112 2: 2
• Reading and writing from files
to fine Java commands Google “java file IO example”
• Paths
absolute /home/desktop/COMP112 and
relative desktop/COMP112
• try { risky code
}
catch { error handling code
}
• Scanners
Introduction
COMP 112 2: 3
• Java file handling uses many Classes including:
1. File
• FileReader
• FileWriter
2. Path
• Look at the API to find specialised file handling.
• Scanners transform a string into a sequence of tokens.
• Frequently the contents of asci files are read using a scanner.
• Scanners implement the interface Iterator<String>
Files
COMP 112 2: 4
• Java will find files in the directory from which the program is
run just by using the file name.
• Java IO design is based on the Decorator pattern
• Thus can lead to unhelpfully complex libraries!
• Beware java NIO (nonblocking) exists as well
• My advice is keep it simple and understand the basics
1. File internal representation of a file
2. FileReader for returning one byte at a time
3. BufferedReader for returning one buffer full at a time
4. Scanner for returning one word (or token) at a time
Streams and conveyer belts.
COMP 112 2: 5
• Stream processing is a common programming idiom based
on manufacturing line:
• Each person works on the part in front of them. Rate of work
is controlled by the conveyer belt.
new Scanner(new FileReader("README.TXT")));
Tokens
Bytes
file
new BufferedReader(new FileReader("README.TXT")));
BufferFull
Bytes
file
Risky code
COMP 112 2: 6
• Any thing that is risk (opening a file that might not exist)
should be performed in a try block try { …} and you need to
catch these errors catch { …} and process them. Even if only
by printing out an error message.
• Uncaught errors crash the program!
COMP 112 2: 7
Example
• From “README.TXT” we build a File then a Scanner.
Build a Scanner
File opening can fail
Create File representation
Failing can throw exceptions
Always Close
Catch IOexceptions
from the try block
Paths and Files
COMP 112 2: 8
• Files are held in a rooted directory tree structure.
• A Complete Path names directories from the root to the file.
/home/users/dstr/Comp112/Lectures/foo.ppt
• A Relative Path names directories from the current location to
the file. If the program is run from /home/users/dstr the files
relative path is Comp112/Lectures/foo.ppt
• Note absolute paths start with a / but relative paths do not.
• Unix and Microsoft computers use paths with different
separators. Unix uses / but Microsoft uses \. Java can use
File.separator and the JVM will automatically adjust
depending on the operating system
File Choosing
COMP 112 2: 9
• When you do not know the name of the file but want to
browse the directories use,
s = new Scanner(new File(UIFileChooser.open()));
in place of,
s =new Scanner(new File("README.TXT")));
DO NOT USE UIFileChooser.open() when you do know the
name of the File! Every year students lose marks in the exam
by using UIFileChooser.open() to open a file of a given name!
Scanner Class
COMP 112 2: 10
• The Scanner class provides all the text input methods from
the UI class. But can take input from Files
s = new Scanner(new File(“README.TXT”));
• The following methods you should recognize
hasNext, hasNextInt, hasNextBoolean, hasNextDouble
next, nextInt, nextBoolean, nextDouble, nextLine
• Scanner can take as input:
1. File
2. Path
3. String
Delimiter
COMP 112 2: 11
• Scanner default delimiter is white space that is: the sequence
of token the scanner reads are separated by white space.
• To change the delimitor use
• s.useDelimiter(",\\s*");
Exercises:
COMP 112 2: 12
• Read a file where each line can have different number of
tokens (words). Output the sequence of tokens and special
token to mark where the lines end.
• Read a file with n integers on each of m lines and output an
n*m array of integers.
File Writing
• Writing text to a file
• All file IO must be in try catch
• Reading and writing files is slow!
COMP 112 2: 13
Buffered File Writing
COMP 112 2: 14
• To save time Buffer your IO
• Note newLine() is platform independent.
• Data may be stored in the buffer and written to the file when
1. the buffer is full
2. file is closed
3. file is flushed
File PrintWriting
COMP 112 2: 15
• Buffered output for text:
• Note println
• PrintWriters are buffered and have automatic line flushing!
Object IO
• Most objects can be serialized to class directory
• class Student implements Serializable {…
• This allows them to written to a file
• and then read back.
COMP 112 2: 16
Overview
COMP 112 2: 17
• My advice is keep it simple and understand the basics
1. File internal representation of a file
2. FileReader/FileWriter for unbuffered reading/writing
BufferedReader/BufferedWriter/PrintWriter for buffered
reading/writing
3. BufferedReader/PrintWriter for reading/writing lines
4. Scanner for reading tokens (words)
5. ObjectInputStream/ObjectOutputStream for
reading/writing Objects
• Use buffered I/O for Reading/Writing large amounts of data.
Data may be lost on failure so flush()buffers.