效率不高的存储桶排序？

Question

I'm trying to use the bucket sort algorithm to sort strings. The task says the runtime should be about 0.05 seconds, but mine takes over 9. Why is mine taking so long, and how can I make it faster. It has about 90000 names in the file. Am I even doing the bucket sort properly?

public static void bucketSortByLength() {
      String[] bucket = new String[14];
      String[] insideBucket;
      int index = 0;
      for(int i = 0; i <= 13; i++)
        bucket[i] = "";
      for(int i = 0; i < numNames; i++)
        bucket[names[i].length()] += names[i] + " ";
      for(int i = 0; i <= 13; i++){
        insideBucket = bucket[i].split("\\s+");
        for(String s : insideBucket)
          names[index++] = s;
      }
    }

Answer 1

Your code is likely slow because it's doing something that can be really inefficient in Java: String values are immutable, so concatenation creates a whole new string each time.

You should use a data type that supports efficient appends for your buckets. Two options that come to mind are StringBuilder and LinkedList<String> . (The latter would have been better when I last used Java everyday, and probably still is.)

Answer 2

As others have stated, this is inefficient because of the string concatenation.

A good rule of thumb:
Whenever you find your self concatenating strings inside a loop, it's best to use a StringBuilder .
See: When to use StringBuilder in Java

This code should give you the performance you are expecting:

public static void bucketSortByLength() {
    StringBuilder[] bucket = new StringBuilder[14];
    int index = 0;
    for(int i = 0; i <= 13; i++)
        bucket[i] = new StringBuilder();
    for(int i = 0; i < numNames; i++)
        bucket[names[i].length()].append(names[i]).append(' ');

    for(int i = 0; i <= 13; i++){
        String[] insideBucket = bucket[i].toString().split("\\s+");
        for(String s : insideBucket)
            names[index++] = s;
    }
}