apriori - Candidate set generation in Apriori algorithm

Question

I am trying to implement Apriori algorithm in Java, and have problems with generating Candidate itemsets. To create candidates for k-itemset I use all combinations of k-1 and 1-itemsets. For example, for Frequent 1-itemset: bread:9, milk:9, coffee:9, sugar:10.

Candidate 2-itemsets generated should be: bread milk, bread coffee, bread sugar, milk coffee, milk sugar, coffee sugar.

But my code returns: bread coffee, bread milk, bread sugar, coffee bread, coffee milk, coffee sugar, milk bread, milk coffee, milk sugar, sugar bread, sugar coffee, sugar milk (all permutations; returns both bread milk and milk bread, however, these two are the same thing).

My code:

public static ArrayList<ArrayList<String>> getCandidates(Map<String, Long> one_itemset_1, Map<String, Long> n_itemset_1){
    ArrayList<ArrayList<String>> tuples = new ArrayList<ArrayList<String>>();


    Map<String, Long> one_itemset = sortbykey(one_itemset_1);
    Map<String, Long> n_itemset = sortbykey(n_itemset_1);

    for(String item : one_itemset.keySet()) {

        for(String item2 : n_itemset.keySet()) {
            if(!item.equals(item2) && !item2.contains(item)) {
                ArrayList<String> singleList = new ArrayList<String>();
                singleList.add(item);
                String item2_sep[] = item2.split(" ");
                for(int i = 0; i < item2_sep.length; i++)
                    singleList.add(item2_sep[i]);
                //singleList.add(item2);
                tuples.add(singleList);
            }
            //index2++;
        }
        //index2 = 0;
        //index1++;
    }

    return tuples;
}

Is there a way to modify this method to get rid of repetitive itemsets? Please advise. Thank you.