0

I have a source HashMap in Java:

HashMap<String, Integer> keyWordFrequencies;

Storing keywords of various length. I want to traverse this HashMap and work out the lengths of the ngrams stored in the String part of the map which defines the text of each keyword.

With this data, I want to populate a target ArrayList of HashMaps:

ArrayList<HashMap<String, Integer>> keywordNgrams;

With the results, where the index of the ArrayList corresponds to the ngram size of a given keyword minus one, i.e. keywordNGrams(0) will receive the unigrams, keywordNGrams(1) will receive the bigrams and so on. But I'm not sure of the necessary syntax. Traversing the source HashMap is easy enough:

Set keyWordFrequenciesSet = keyWordFrequencies.entrySet();
Iterator keyWordFrequenciesIterator = keyWordFrequenciesSet.iterator();
while(keyWordFrequenciesIterator.hasNext()) {
   Map.Entry m = (Map.Entry) keyWordFrequenciesIteratorIterator.next();
   int ngramLength = String_Utils.getLengthOfNgram(m.getKey().toString());

   Add element to keywordNgrams?

But adding the element to the target ArrayList of HashMap is confusing me. I have tried:

keywordNgrams.add(ngramLength, m);

And various alternatives but to no avail. m should be an element of a HashMap, not a HashMap in itself. Can anyone suggest where I am wrong?

Ideally, I would like to traverse the source HashMap keyWordFrequencies once, and the keywordNgrams ArrayList is initialised to the largest possible ngram size to start with.

Improve this question
5
  • Before you begin, do you start out already knowing what the greatest ngram-size will be? That is, do you know what the final value of keywordNgrams.size() will be? If so, then you're best off pre-populating keywordNgrams to that size; if not, then whenever you encounter a larger ngram than you've seen before, you'll need a loop to expand keywordNgrams to the desired size. Commented Sep 12, 2012 at 19:09
  • yes, I do know this. The greatest length of an ngram should be 5. Commented Sep 12, 2012 at 19:10
  • keywordNgrams.add(ngramLength, m); will not work because m is a Map.Entry instance, not an instance of a HashMap. Are you sure you want an ArrayList<HashMap> and not an ArrayList<Map.Entry>? With an ArrayList<HashMap>, you will be creating a list of 5 elements, with each element being the exact same HashMap which I doubt is what you intend. Commented Sep 12, 2012 at 19:16
  • And then access the key and value via m itself? Sounds promising. Commented Sep 12, 2012 at 19:19
  • Thanks to the respondents: I will try these out in the morning and see. Commented Sep 12, 2012 at 19:23

2 Answers 2

Reset to default
1

Since you're dealing with an ArrayList of size 5, I would suggest that when you initialize your ArrayList, do so by adding a new instance of a HashMap at each index. Something like this:

ArrayList<HashMap<String, Integer>> keywordNgrams = new ArrayList<HashMap<String, Integer>> ();

for(int index = 0; index < 5; index++){
  keywordNgrams .put(index, new HashMap<String, Integer>());
}

In order to add elements in your ArrayList, here's what you've to do:

  • Access the specific HashMap for a specified 'n'-gram. This you can do using get(int index) on the ArrayList
  • You would then add the element your returned HashMap and then again do a set(int index, E element) of the same HashMap to your keywordNgrams ArrayList.

A sample code might be something like this:

HashMap<String, Integer> returnedMap = keywordNgrams.get(index); //where index is the position in the list;
returnedMap.put(key, value); //where key & value is the information that you would want to add to your HashMap
keywordNgrams.set(index, returnedMap);
Improve this answer
Sign up to request clarification or add additional context in comments.

Comments

0

Since you know the greatest ngram-size, I recommend prepopulating keywordNgrams:

List<Map<String, Integer>> keywordNgrams =
    new ArrayList<Map<String, Integer>>();
for(int i = 0; i < 5; ++i)
    keywordNgrams.add(new HashMap<String, Integer>());

Then you can write:

for(final String keyword : keyWordFrequencies.keySet())
    keywordNgrams.get(String_Utils.getLengthOfNgram(keyword) - 1)
            .put(keyword, keyWordFrequencies.get(keyword));
Improve this answer

4 Comments

This will populate the target keywordNgrams HashMaps with the correct ngrams but without the frequencies of them from the source keyWordFrequencies. But by removing the if...else with: map.put(key.toString(), ((Integer) keywordFrequencies.get(key.toString())).intValue()); it seems to be alright. Many thanks.
@MrMorgan: Ah, I see. I couldn't figure out why your original map was from String to Integer, but I get it now. I'll update my answer accordingly.
@MrMorgan: I've updated my answer. As you can see, you shouldn't need to use key.toString(), and you shouldn't need to cast to Integer, as long as you preserve the types to begin with: since keyWordFrequencies has type HashMap<String, Integer>, the casts should already be taken care of for you by the compiler.
FAO ruakh: I've tried your modified code and it appears to work nicely. To clarify things slightly, I need to use dedicated target HashMaps to record the number of times each ngram occurs in in the original HashMaps, i.e. so keywordNgrams(0) stores the unigrams, keywordNgrams(1) stores the bigrams and their frequencies, and keywordNgrams(2) will store the trigrams and so on. I suppose an alternative would just be to have one really long target HashMap storing all of the uni, bi and tri grams together but that's possibly less flexible. Thanks for your help.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.