Character encoding. In short, it describes how bytes are converted into characters. The default character encoding is determined by your operating system, and it may not be the actual encoding of the file. If it isn't then FileWriter is not a good option.
On a more general level, it means that the FileWriter class abstracts away those two parameters (character encoding and buffer size) - it uses default values that your code can't change. Should you do need to change them, then using an OutputStreamWriter (for the encoding) and a BufferedWriter (for the buffer size) in conjunction with a FileOutputStream are called for. The javadocs of those other classes provide more information about what exactly they do.
Joined: Oct 31, 2009
Thank you for the answers.
I thought I had figured things out after getting your warning about FileWriter and reading the docs, but I am apparently still missing something.
I have a text-file which was likely created on a Windows machine. According to the "file" tool on my linux machine, its encoding is "ISO-8859".
I extract certain strings from this file and write those to a new file. I have done the following, where inFile is the File object for the input file,
and outFile is the File object for the output file.
I write to osw using
Now when I open outFile in a texteditor a questionmark appears for instance where there ought to be a character with an accent egu and checking the type of
outFile with "file" again, shows that it is ASCII despite having chosen "ISO-8859-1" in the OutputStreamWriter constructor .
Joined: Oct 31, 2009
Apologies for posing my last question. Hope you havent wasted too much time thinking about it.
This was for a term-project and we were giving code to load some data. On linux machines some bugs occured. This was caused by the provided code using FileWriter. The above code does fix this. I had overseen that the writer was reassigned after initialization apparently for the sole silly purpose of skipping the first line in the file !