一般来说int代表一个数字,但是如果利用每一个位 ,则可以表示32个数字 ,在数据量极大的情况下可以显著的减轻内存的负担。我们就以int为例构造一个bitmap,并使用其来解决一个简单的问题:求两个数组的交集
先实现一个bitmap
/** * @Description: * @author: zhoum * @Date: 2020-01-23 * @Time: 10:49 */ public class BitMap { private int[] sign = {0x00000001,0x00000002,0x00000004,0x00000008,0x00000010,0x00000020,0x00000040,0x00000080,0x00000100,0x00000200,0x00000400,0x00000800,0x00001000,0x00002000,0x00004000,0x00008000, 0x00010000,0x00020000,0x00040000,0x00080000,0x00100000,0x00200000,0x00400000,0x00800000,0x01000000,0x02000000,0x04000000,0x08000000,0x10000000,0x20000000,0x40000000,0x80000000}; private int[] arr ; private int capacity; public BitMap(int capacity) { validate(capacity); this.capacity = capacity; this.arr = new int[(capacity>>5)+1]; } public void put(int k){ if ( k > capacity ){ throw new RuntimeException("k is greater than capacity"); } validate(k); int index = k >> 5 ;//当前数字应该存放的bucket索引 arr[index] = arr[index]|sign[k & 31]; } private void validate(int k){ if ( k <= 0 ){ throw new IllegalArgumentException(" capacity must be greater than zero"); } } public int[] getMixed(BitMap bitMap){ int length = Math.min(bitMap.arr.length,this.arr.length); int[] other = new int[length],me = new int[length]; System.arraycopy(bitMap.arr,0,other,0,length); System.arraycopy(this.arr,0,me,0,length); //借用集合的无固定大小来构建最后数组 List<Integer> result = new ArrayList<>(); for (int i = 0; i < length; i++) { int k= other[i] & me[i]; for (int j = 1; j <= 32; j++) { if ( ((k>>j)&1) == 1 ){ result.add((i<<5)+j); } } } if ( result.size() == 0 ){ return null; }else { int[] rs = new int[result.size()]; for (int i = 0; i < result.size(); i++) { rs[i] = result.get(i); } return rs; } } }
写一个main方法试验下
public static void main(String[] args) {
BitMap bitMap = new BitMap(1000){{ put(248); put(5); put(9); put(12); put(6); put(13); put(963); }}; BitMap bitMap1 = new BitMap(1000){{ put(248); put(15); put(13); put(963); put(5); put(6); put(9); }}; int[] mixed = bitMap.getMixed(bitMap1); System.out.println(Arrays.toString(mixed)); }
得到有序结果