Streams
Hi.
I am wondering what the difference between byte and character streams are.
One thing I already know:
Byte Streams can read one single byte at a time while Character Streams read two bytes at a time (the number of bytes that compose the primitive type char).
Altough we have this great difference, I have two source files. the first using Byte streams and the second using character streams, both just read an input file and writes it out to another file.
They do the same thing nothing goes wrong. So what's the difference?
Thanks
A byte stream is one dang byte after another.
A character stream is one dang character after another.
Now when it comes to *encoding* character streams as byte streams or
*decoding* byte streams to character streams, then you might find a character
corresponds to one, two or even more bytes, depending on the encoding.
As for what you are doing, copying a file, unless you are changing the character encoding, doing the copy byte-wise makes more sense.
Character streams also apply a character encoding to the stream, so that the data is converted from bytes to characters correctly. If you read and write the stream using the same encoding, then you won't see any difference from using a byte stream. However, try reading it with one encoding, and writing it with another and see what difference it makes. The rule of thumb is, if it's text, use a character stream, if it's data, use a byte stream.
Hi bigdaddy.
Well I got what you mean.
Actually the prim. type Char in java follows Unicode Conventions, every char is made of 16bits(2bytes). I don't undestand how A byte stream can read half(8bits) and a half(8bits) of character and succeds.
So there is no difference between reading a file using byte or character strings?
It sounds strange to me.
are you talking about input or output streams?Are you talking about the difference between streams and readers/writers?
What do you mean by data stream? Would it be binary ?Thanks
Input and output streams:)
> Hi bigdaddy.
>
> Well I got what you mean.
> Actually the prim. type Char in java follows Unicode
> Conventions, every char is made of 16bits(2bytes). I
> don't undestand how A byte stream can read
> half(8bits) and a half(8bits) of character and
> succeds.
>
> So there is no difference between reading a file
> using byte or character strings?
> It sounds strange to me.
There is a difference between reading data as bytes or characters. A char stream applies a character encoding. A byte stream doesn't.
Read through this IO tutorial, it explains things:
http://java.sun.com/docs/books/tutorial/essential/io/
It is exactly the trail I am following.Maybe such explanation is in the end of the trail. Thnaks
> Well I got what you mean.
> Actually the prim. type Char in java follows Unicode
> Conventions, every char is made of 16bits(2bytes). I
> don't undestand how A byte stream can read
> half(8bits) and a half(8bits) of character and
> succeds.
I'm not sure what you mean here. But consider how streams are connected
in order to read text from a file:
InputStream is= new FileInputStream(file);
Reader r = new InputStreamReader(is, "thecharsetname");
FileInputStream produces a stream of bytes, and InputStreamReader
converts this into a stream of chars. How many bytes are required
to generate a char depend on the charset encoding. For example, the
default windows encoding and the default UNIX encoding both have
one byte per char.
> So there is no difference between reading a file
> using byte or character strings?
> It sounds strange to me.
There is a difference -- reading a file as a character stream involves
decoding the byte stream into a char stream.
> Actually the prim. type Char in java follows Unicode
> Conventions, every char is made of 16bits(2bytes). I
> don't undestand how A byte stream can read
> half(8bits) and a half(8bits) of character and
> succeds.
This used to be true in earlier versions of Unicode. Unicode is no longer a 16-bit character set.
In newer versions of Java Strings are not bound to the 16-bit characters; however, the Character class is still bound due to its encapsulated value being a primitive char.
See the Javadocs for Character pretty detailed explanation.
jbisha at 2007-7-12 20:48:49 >
Thanks BigIt is clearer now:)
One more question BigIf I am not mistaken, char streams convert bytes to the local character set and from the local character set to bytes.In other words ... I can specify a characterset or not, Is it true?
> One more question Big
> If I am not mistaken, char streams convert bytes to
> the local character set and from the local character
> set to bytes.
> In other words ... I can specify a characterset or
> not, Is it true?
Character streams use a default encoding if you don't specify one. That default may or may not match the actual encoding of the file, so it's always a good idea to specify.
That is it. Now I know Why the examples I have here both worked using either byte and character streams
The contents of the file is US format. there are no special characters such as in portuguese (my language).
If I had tried to read a brazilian portuguese file, I would certainly have problems because the default charset of my linux is US.
tHANK YOU A LOT
