Does the Scanner class load the entire file into memory at once?

问题

I often use the Scanner class to read files because it is so convenient.

      String inputFileName;
      Scanner fileScanner;

      inputFileName = "input.txt";
      fileScanner = new Scanner (new File(inputFileName));

My question is, does the above statement load the entire file into memory at once? Or do subsequent calls on the fileScanner like

      fileScanner.nextLine();

read from the file (i.e. from external storage and not from memory)? I ask because I am concerned about what might happen if the file is too huge to be read into memory all at once. Thanks.

回答1:

If you read the source code you can answer the question yourself.

It appear that the implementation of the Scanner constructor in question shows:

public Scanner(File source) throws FileNotFoundException {
        this((ReadableByteChannel)(new FileInputStream(source).getChannel()));
}

Latter this is wrapped into a Reader:

private static Readable makeReadable(ReadableByteChannel source, CharsetDecoder dec) {
    return Channels.newReader(source, dec, -1);
}

And it is read using a buffer size

private static final int BUFFER_SIZE = 1024; // change to 1024;

As you can see in the final constructor in the construction chain:

private Scanner(Readable source, Pattern pattern) {
        assert source != null : "source should not be null";
        assert pattern != null : "pattern should not be null";
        this.source = source;
        delimPattern = pattern;
        buf = CharBuffer.allocate(BUFFER_SIZE);
        buf.limit(0);
        matcher = delimPattern.matcher(buf);
        matcher.useTransparentBounds(true);
        matcher.useAnchoringBounds(false);
        useLocale(Locale.getDefault(Locale.Category.FORMAT));
    }

So, it appears scanner does not read the entire file at once.

回答2:

From reading the code, it appears to load 1 KB at a time by default. The size of the buffer can increase for long lines of text. (To the size of the longest line of text you have)

回答3:

You're better off going with something like BufferedReader with a FileReader for large files. A basic example can be found here.

回答4:

In ACM Contest the fast read is very important. In Java we found found that use something like that is very faster...

    FileInputStream inputStream = new FileInputStream("input.txt");
    InputStreamReader streamReader = new InputStreamReader(inputStream, "UTF-8");
    BufferedReader in = new BufferedReader(streamReader);
    Map<String, Integer> map = new HashMap<String, Integer>();
    int trees = 0;
    for (String s; (s = in.readLine()) != null; trees++) {
        Integer n = map.get(s);
        if (n != null) {
            map.put(s, n + 1);
        } else {
            map.put(s, 1);
        }
    }

The file contains, in that case, tree names...

Red Alder
Ash
Aspen
Basswood
Ash
Beech
Yellow Birch
Ash
Cherry
Cottonwood

You can use the StringTokenizer for catch any part of line that your want.

We have some errors if we use Scanner for large files. Read 100 lines from a file with 10000 lines!

A scanner can read text from any object which implements the Readable interface. If an invocation of the underlying readable's Readable.read(java.nio.CharBuffer) method throws an IOException then the scanner assumes that the end of the input has been reached. The most recent IOException thrown by the underlying readable can be retrieved via the ioException() method.

tells in the API

Good luck!

来源：https://stackoverflow.com/questions/10336478/does-the-scanner-class-load-the-entire-file-into-memory-at-once

标签

java

java.util.scanner