read() method working in byte stream and character stream classes

[sonali raina]

what is the exact difference between read() method of FileInputStream and FileReader class?I tried to read and write a character with unicode valueFfE6 i.e. 16 bit,read() of both the classes was able to read the 16 bit char successfully.Then why is it said that read() of FileInputStream class holds a byte value?

[Richard Tookey]

sonali raina wrote:I tried to read and write a character with unicode valueFfE6 i.e. 16 bit,read() of both the classes was able to read the 16 bit char successfully.

Sorry but that can't be the whole story since, except for the EOF value, the range of values you can obtain from FileInputStream.read() is 0 to 0xFF inclusive ! Please show the code you used to get the value 0xF1E6 from FileInputStream.read() .

[sonali raina]

I used the code from http://docs.oracle.com/javase/tutorial/essential/io/charstreams.html
http://docs.oracle.com/javase/tutorial/essential/io/bytestreams.html
but in the xanadu.docx file I copied the characters from character-map(in accessories).In character map select character set:Unicode and try to use characters with larger unicode value(say FE22)
Also use docx file because simple txt file doesnot support unicode encoding.

[Richard Tookey]

Sorry but I really don't understand what code you used to get byte values of 0xFF22 since nothing in those references can possibly generate byte values out of the range 0 to 0xFF !

[sonali raina]

Thank you for the reply Richard Tookey
I have used exactly the same code(CopyBytes) which is given in http://docs.oracle.com/javase/tutorial/essential/io/bytestreams.html link.
That is I am trying to copy characters from "xanadu.docx" to "outagain.docx".
But the real confusion is that even chinese characters (like "full width won sign ￦"and "full width Yen sign ¥") are also read and copied successfully from xanadu to outagain; which i think should not happen because chinese characters are represented in 16 bits in unicode encoding and FileInputStream.read() reads only 8 bits.

[Richard Tookey]

sonali raina wrote:
I have used exactly the same code(CopyBytes) which is given in http://docs.oracle.com/javase/tutorial/essential/io/bytestreams.html link.
That is I am trying to copy characters from "xanadu.docx" to "outagain.docx".

Sorry but that code copies bytes and not characters and nothing in it prints the bytes being transferred. Since in a file, depending on the character encoding, all characters are represented as one or my bytes I don't understand your problem. Copying the bytes from a file to a file is in effect copying the characters. It is only within memory where one actually has a 'char' primitive and the conversion of bytes to a char is not involved at all in the code you cite.

I suspect someone else will have to respond since I really don't understand your problem.