Reading unicode character in java

Question

I'm a bit new to java, When I assign a unicode string to

  String str = "\u0142o\u017Cy\u0142";
  System.out.println(str);

  final StringBuilder stringBuilder = new StringBuilder();
  InputStream inStream = new FileInputStream("C:/a.txt");
  final InputStreamReader streamReader = new InputStreamReader(inStream, "UTF-8");
  final BufferedReader bufferedReader = new BufferedReader(streamReader);
  String line = "";
  while ((line = bufferedReader.readLine()) != null) {
      System.out.println(line);
      stringBuilder.append(line);
  }

Why are the results different in both cases the file a.txt also contains the same string. but when i print output of the file it prints z\u0142o\u017Cy\u0142 instead of the actual unicode characters. Any idea how do i do this if i want to file content also to be printed as string is being printed.

Answer 1

I think its just "UTF8" not "UTF-8".

Here I saw it: Source

Answer 2

It sounds as though your file literally contains the text z\u0142o\u017Cy\u014, i.e. has Unicode escape sequences in it.

There's probably a library for decoding these but you could do it yourself - according to the Java Language Specification an escape sequence is always of the form \uxxxx, so you could get the 4-digit hex value xxxx for the character, convert it to an integer with Integer.parseInt, convert it to a character and finally replace the whole \uxxxx sequence with the character.

Answer 3

Your code should be correct, but I guess that the file "a.txt" does not contain the Unicode characters encoded with UTF-8, but the escaped string "\u0142o\u017Cy\u0142".

Please check if the text file is correct, using an UTF-8 aware editor such as recent versions of Notepad or Notepad++ on Windows. Or edit it with your favorite hex editor - it should not contain backslashes.

I tried it with "€" as UTF-8-encoded content of the file and it gets printed correctly. Note that not all Unicode characters can be printed, depending on your terminal encoding (really a hassle on Windows) and font.

Answer 4

Java interprets unicode escapes such as your \u0142 that are in the source code as if you had actually typed that character (latin small letter L with stroke) into the source. Java does not interpret unicode escapes that it reads from a file.

If you take your String str = "\u0142o\u017Cy\u0142"; and write it to a file a.txt from your Java program, then open the file in an editor, you'll see the characters themselves in the file, not the \uNNNN sequence.

If you then take your original posted program and read that a.txt file you should see what you expected.

Answer 5

So, you want to unescape unicode codepoints? There is no public API available for this. The java.util.Properties has a loadConvert() method which does exactly this, but it's private. Check the Java source for the case you'd like to reuse this. It's doing the conversion by simple parsing. I wouldn't use regex for this since this is too error prone in very specific circumstances.

Or you should probably after all be using java.util.Properties or its i18n counterpart java.util.ResourceBundle with a .properties file instead of a plain .txt file.

See also:

• Properties API tutorial
• Internationalization tutorial - ResourceBundle.

Answer 6

You have used FileInputStream and is a byte code reader not character reader. Try using FileReader instead

something like:

BufferedReader inputStream = new BufferedReader(new FileReader("C:/a.txt"));

then you can use the line oriented I/O BufferedReader to read each line. FileInputREader is a low level I/O that you should avoid. You're writing the characters to your file not the bytes, the best approach is to use character streams. for wrinting and reading unless you need to write bytes/binary data.