如何避免时间复杂度O(n^2)？

Question

I have this following piece of code:

for(ArticleBasketInBean basketBean : bean.getBasket()) {
    for(ArticleDTO article : dto.getArticleList()) {
        if(basketBean.getArticleReference().equals(article.getArticleReference())) {
            article.setAddedToBasket(true);
        }
    }
}

Clearly the time complexity of the above operation is O(n^2). For this case articleReference is unique. So the list returned by bean.getBasket() has no duplicate articleReference , also this is true for the list returned by dto.getArticleList() .

I want to avoid this nested iteration and want to write faster code. How can I do that?

Answer 1

Use java.util.HashSet to cache one of the sets of references, assuming the sets are not TOO large, of course. With a good hash function, this should bring you back to O(n).

Answer 2

A simple break will save you.

    for(ArticleBasketInBean basketBean : bean.getBasket()) {
    for(ArticleDTO article : dto.getArticleList()) {
        if(basketBean.getArticleReference().
                                  equals(article.getArticleReference())) {
            article.setAddedToBasket(true);
            break;
        }
    }
}

Answer 3

If you don't have too many articles/baskets and as you say the references are unique you can do something like this; let's call R the type of the reference, A the type of articles, B the type of baskets:

// since all references are unique we can use that, but see below
Map<R, A> mappedArticles = new IdentityHashMap<>();

// inject articles based on references in the map, then

A article;

for (B basket: bean.getBasket())
    article = mappedArticles.get(basket.getArticleReferences());
    if (article != null)
        article.setAddedToBasket(true);

// Full list of articles is in the map's .values()

Notes:

• note the use of an identity hashset; you may want to implement equals/hashcode for the references instead (or maybe the type you use has them already implemented for you);
• your map may be filled as articles "flow by", in this case you can create it only once (make it thread safe if so!).

Answer 4

This is where hashMap Data Structure comes to rescue. It's not any rocket science, simply:

an object with key, value pair

The complexity will reduce from O(n2) to O(n), which is better.