当元素超过其大小的 1/2 时调整数组大小

Question

当元素数 N 大于 m/2 时，我试图调整数组的大小，m 是数组的初始大小，但它不起作用，我不明白为什么。这个数组应该像一个哈希表一样工作，所以我在每次插入之前都有一个哈希函数，在调整大小之后我想再次插入每个元素，并使用新的哈希值（m 值已更改）。这是错误：

Exception in thread "main" java.lang.OutOfMemoryError: Java heap space
    at JumpHashing.resize(JumpHashing.java:55)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)
    at JumpHashing.hashing(JumpHashing.java:40)
    at JumpHashing.resize(JumpHashing.java:61)
    at JumpHashing.put(JumpHashing.java:50)

问题显然是调整大小，没有它（少于 23 个元素）它可以工作。

m的初始大小是23，这是实际代码（从algs4读取文件的类“In”）：

public class JumpHashing{
    private int m;
    private int[] hashTable; 
    private static int id;
    private int N;

    public JumpHashing(){
        m = 23;
        hashTable = new int[m];
        N = 0;
    }

    public void hashing(int value) {
            int key = (value*11)%m;
            put(key, value);
    }

    public void put(int key, int value) {
        if(N <m/2) {
            hashTable[key] = value;
            N++;
        } else {
            m=m*2;
            N=0;
            resize(m);
        }
    }

    public void resize(int m) { 
        int[] temp = new int[m];
        for(int i=0; i<hashTable.length; i++) {
            temp[i] = hashTable[i];
        }
        hashTable = new int[m];
        for(int i=0; i<temp.length; i++) {
            hashing(temp[i]);
        }
    }

    public static void main(String[] args) {
        JumpHashing hashT1 = new JumpHashing();

        In file = new In("random.txt");
        while(file.hasNextLine()) {
            int value = Integer.parseInt(file.readLine());
            hashT1.hashing(value);
        }   
        for(int j=0; j<hashT1.hashTable.length; j++) {
            StdOut.println("Key: "+j+" Value: "+hashT1.hashTable[j]);
        }
    }
}

Answer 1

您最终会反复调用resize，直到内存用完。问题出在这个函数中：

    public void resize(int m) { 
        int[] temp = new int[m];  // <-- this is the new double-size of m
        for(int i=0; i<hashTable.length; i++) {
            temp[i] = hashTable[i];
        }
        hashTable = new int[m];
        for(int i=0; i<temp.length; i++) {  // <-- here we go too far
            hashing(temp[i]);
        }
    }

您的第二个循环遍历全新的“m”大小数组，而不是原始 m/2 大小数组。循环的一半加一，您的N 将再次大于m/2，并且每次发生这种情况时都会调用 resize。

这是您在该功能中应该具备的：

public void resize(int m) {
    int[] oldHash = hashTable;
    hashTable = new int[m];
    for(int i=0; i<oldHash.length; i++) {
        if (oldHash[i] != 0) {     // <-- don't hash empty slots
            hashing(oldHash[i]);
        }
    }
}

这也提高了性能，因为您只循环一次并且不超过 m/2 次。