Question

How can I find commonly occurring sequences in a string in Java?

The string is a long sequence of digits and I want to find the most commonly occurring sequences of digits.

Answer 1

I guess that depends on how long the sequences are that you are looking for.

What I would do is use a Guava Multiset, iterate over the sequence, write all subsequences to the Multiset and sort that by occurrence. Here’s a sample implementation:

public static String getMostFrequentSequence(final String input, final int patternLength) {
    final Multiset<String> multiset = HashMultiset.create();
    final int length = patternLength < 0 ? input.length() : Math.min(patternLength, input.length());
    for (int l = 2; l < length; l++) {
        for (int o = 0; o < input.length() - l; o++) {
            multiset.add(input.substring(o, o + l));
        }
    }
    return Ordering.from(new Comparator<Entry<String>>() {
        public int compare(final Entry<String> o1, final Entry<String> o2) {
            return
            ComparisonChain.start()
            .compare(o1.getCount(), o2.getCount())
            .compare(o1.getElement(), o2.getElement())
            .result();
        }
    }).max(multiset.entrySet()).getElement();
}

And about performance: This test method takes about a second on my machine for unlimited length and about 25 milliseconds when I limit the pattern length to 12 chars

public static void main(final String[] args) throws Exception {
    final StringBuilder sb = new StringBuilder();
    final Random random = new Random();
    for (int i = 0; i < 1000; i++) {
        sb.append(random.nextInt(10));
    }
    final long t1 = System.currentTimeMillis();
    final String input = sb.toString();
    System.out.println(input);
    System.out.println(getMostFrequentSequence(input, -1));
    System.out.println(System.currentTimeMillis() - t1);
    final long t2 = System.currentTimeMillis();
    System.out.println(getMostFrequentSequence(input, 12));
    System.out.println(System.currentTimeMillis() - t2);
}