CharsRefIntHashMap并不比HashMap<String, Integer>快
我模仿lucene的BytesRef写了一个CharsRefIntHashMap,实測效果并不如HashMap<String, Integer>。代码例如以下:
package com.dp.arts.lucenex.utils;
import org.apache.lucene.util.CharsRef;
public interface CharsRefIntMap
{
public static
abstract class CharsRefIntEntryAccessor {
public abstract
void access(char[] arr,
int offset, int length,
int value);
}
public void incKey(CharsRef key);
public void incKey(CharsRef key,
int add);
public void incKey(char[] arr,
int offset, int length);
public void incKey(char[] arr,
int offset, int length,
int add);
public int get(CharsRef key);
public int get(CharsRef key,
int no_entry_value);
public int get(char[] arr,
int offset, int length);
public int get(char[] arr,
int offset, int length,
int no_entry_value);
public int size();
public void forEach(CharsRefIntEntryAccessor accesor);
}
import java.util.Arrays;
import org.apache.lucene.util.CharsRef;
import com.dp.arts.lucenex.utils.CharsRefIntMap.CharsRefIntEntryAccessor;
public class CharsRefIntHashMap implements CharsRefIntMap
{
public static final int DEFAULT_CAPACITY = 16;
private char[][] arrs;
private int[] offsets;
private int[] lengths;
private int[] ords;
private int[] values;
private int hashSize;
private int halfHashSize;
private int hashMask;
private int count;
public CharsRefIntHashMap() {
this(DEFAULT_CAPACITY);
}
public CharsRefIntHashMap(int capacity) {
assert capacity > 0 && ( (capacity & (capacity - 1)) == 0);
arrs = new char[capacity][];
offsets = new int[capacity];
lengths = new int[capacity];
ords = new int[capacity];
values = new int[capacity];
Arrays.fill(ords, -1);
hashSize = capacity;
halfHashSize = (capacity >>> 1);
hashMask = capacity - 1;
}
@Override
public void incKey(CharsRef key) {
int code = charsHashCode(key.chars, key.offset, key.length);
incKey(key.chars, key.offset, key.length, code, 1);
}
@Override
public void incKey(CharsRef key, int add) {
int code = charsHashCode(key.chars, key.offset, key.length);
incKey(key.chars, key.offset, key.length, code, add);
}
@Override
public void incKey(char[] arr, int offset, int length) {
int code = charsHashCode(arr, offset, length);
incKey(arr, offset, length, code, 1);
}
@Override
public void incKey(char[] arr, int offset, int length, int add) {
int code = charsHashCode(arr, offset, length);
incKey(arr, offset, length, code, add);
}
private void incKey(char[] arr, int offset, int length, int code, int add) {
int pos = (code & hashMask);
int e = ords[pos];
while (e != -1 && !charsEquals(arrs[e], offsets[e], lengths[e], arr, offset, length)) {
final int inc = ((code >> 8) + code) | 1;
code += inc;
pos = (code & hashMask);
e = ords[pos];
}
if (e == -1) {
// new entry.
arrs[count] = arr;
offsets[count] = offset;
lengths[count] = length;
values[count] = add;
ords[pos] = count;
++count;
if (count == halfHashSize) {
rehash((hashSize << 1));
}
} else {
values[e] += add;
}
}
private void rehash(int newSize) {
char[][] newArrs = new char[newSize][];
int[] newOffsets = new int[newSize];
int[] newLengths = new int[newSize];
int[] newValues = new int[newSize];
System.arraycopy(arrs, 0, newArrs, 0, halfHashSize);
System.arraycopy(offsets, 0, newOffsets, 0, halfHashSize);
System.arraycopy(lengths, 0, newLengths, 0, halfHashSize);
System.arraycopy(values, 0, newValues, 0, halfHashSize);
final int[] newOrds = new int[newSize];
Arrays.fill(newOrds, -1);
final int newHashMask = newSize - 1;
for (int i = 0; i < hashSize; ++i) {
int e0 = ords[i];
if (e0 != -1) {
char[] arr = newArrs[e0];
int offset = newOffsets[e0];
int length = newLengths[e0];
int code = charsHashCode(arr, offset, length);
int pos = code & newHashMask;
while (newOrds[pos] != -1) {
final int inc = ((code >> 8) + code) | 1;
code += inc;
pos = code & newHashMask;
}
newOrds[pos] = e0;
}
}
ords = newOrds;
arrs = newArrs;
offsets = newOffsets;
lengths = newLengths;
values = newValues;
hashSize = newSize;
halfHashSize = (newSize >> 1);
hashMask = newHashMask;
}
public int charsHashCode(char[] chars, int offset, int length) {
final int prime = 31;
int result = 0;
final int end = offset + length;
for (int i = offset; i < end; i++) {
result = prime * result + chars[i];
}
return result;
}
public boolean charsEquals(char[] lhsArr, int lhsOffset, int lhsLength, char[] rhsArr, int rhsOffset, int rhsLength) {
if (lhsLength == rhsLength) {
int otherUpto = rhsOffset;
final int end = lhsOffset + lhsLength;
for (int upto = lhsOffset; upto < end; upto++, otherUpto++) {
if (lhsArr[upto] != rhsArr[otherUpto]) {
return false;
}
}
return true;
} else {
return false;
}
}
@Override
public int get(CharsRef key) {
return get(key.chars, key.offset, key.length, 0);
}
@Override
public int get(CharsRef key, int no_entry_key) {
return get(key.chars, key.offset, key.length, no_entry_key);
}
@Override
public int get(char[] arr, int offset, int length) {
return get(arr, offset, length, 0);
}
@Override
public int get(char[] arr, int offset, int length, int no_entry_key) {
int code = charsHashCode(arr, offset, length);
int pos = (code & hashMask);
int e = ords[pos];
while (e != -1 && !charsEquals(arrs[e], offsets[e], lengths[e], arr, offset, length)) {
final int inc = ((code >> 8) + code) | 1;
code += inc;
pos = (code & hashMask);
e = ords[pos];
}
return e == -1 ? no_entry_key : values[e];
}
@Override
public void forEach(CharsRefIntEntryAccessor accessor) {
for (int i = 0; i < hashSize; ++i) {
int pos = ords[i];
if (pos != -1) {
accessor.access(arrs[pos], offsets[pos], lengths[pos], values[pos]);
}
}
}
@Override
public int size() {
return count;
}
// for test only.
public int hashSize() {
return hashSize;
}
}
package com.dp.arts.lucenex.utils;
import java.util.HashMap;
import java.util.Random;
import
org.apache.lucene.util.CharsRef;
public class CharsRefIntHashMapBenchmark
{
private static Random
randGen = null;
private
static char[]
numbersAndLetters =
null;
static {
randGen =
new Random();
numbersAndLetters = ("0123456789abcdefghijklmnopqrstuvwxyz" +
"0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ").toCharArray();
}
private static
final String randomString(int length) {
if (length < 1) {
return
null;
}
char [] randBuffer =
new char[length];
for (int i=0; i<randBuffer.length; i++) {
randBuffer[i] = numbersAndLetters[randGen.nextInt(71)];
}
return new String(randBuffer);
}
public static
void main(String[] args) {
final int MAX = 100000;
String[] strs = new String[10000];
int[] values =
new int[MAX];
for (int i = 0; i < 10000; ++i) {
strs[i] = randomString(randGen.nextInt(10) + 1);
}
for (int i = 0; i < MAX; ++i) {
values[i] = randGen.nextInt(10000);
}
char[][] arrs =
new char[MAX][];
int offsets[] =
new int[MAX];
int counts[] =
new int[MAX];
for (int i = 0; i < MAX; ++i) {
String s = strs[values[i]];
arrs[i] = StringMisc.toCharArray(s);
offsets[i] = StringMisc.getOffset(s);
counts[i] = StringMisc.getCount(s);
}
long start = System.currentTimeMillis();
CharsRefIntHashMap map = new CharsRefIntHashMap();
for (int j = 0; j < 100; ++j) {
for (int i = 0; i < MAX; ++i) {
map.incKey(arrs[i], offsets[i], counts[i]);
}}
System.err.println("CharsRefIntHashMap time elapsed: " + (System.currentTimeMillis() - start) +
"ms.");
start = System.currentTimeMillis();
HashMap<String, Integer> oldMap = new HashMap<String, Integer>();
for (int j = 0; j < 100; ++j) {
for (int i = 0; i < MAX; ++i) {
String s = strs[values[i]];
Integer v = oldMap.get(s);
if (v ==
null) {
v = new Integer(1);
oldMap.put(s, v);
} else {
v += 1;
}
}}
System.err.println("Origin string map time elapsed: " + (System.currentTimeMillis() - start) +
"ms.");
}
}
CharsRefIntHashMap并不比HashMap<String, Integer>快的更多相关文章
- string integer == equals 转
java中的数据类型,可分为两类: 1.基本数据类型,也称原始数据类型.byte,short,char,int,long,float,double,boolean 他们之间的比较,应用双等号(== ...
- String,Integer,int类型之间的相互转换
String, Integer, int 三种类型之间可以两两进行转换 1. 基本数据类型到包装数据类型的转换 int -> Integer (两种方法) Integer it1 = new I ...
- Junit + String/Integer/ArrayList/HashMap/TreeMap 基本使用Demo
package JavaTest.test; import java.util.ArrayList; import java.util.HashMap; import java.util.List; ...
- List<Map<String, Integer>> 同key的value全部累加合并
public static void main(String[] args){ List<Map<String,Object>> list1 = new ArrayList&l ...
- Delphi out 参数 string Integer
http://www.delphibasics.co.uk/RTL.asp?Name=Out http://stackoverflow.com/questions/14507310/whats-the ...
- string(Integer)类的equals和==区别和联系（验证密码的时候用得到）
“==”在八种原始数据类型中,判断的是两边的值是否相等.对于对象类型来说,判断的是内存地址,所以为true所满足的条件就是两边的引用指向同一个对象. 比如String s1 = "abcde ...
- java方法笔记： split,instanceof，StringBuilder（），append()，string ,Integer,Math，正则表达式
http://swiftlet.net/archives/709 如果分隔符在前方,会显示“”,在尾部不显示: *|做分割,需要加\\ API文档如下 1.instanceof 查看前者是否是后者的实 ...
- Java String Integer转换 练习：编程求字符串“100”和“150”按十进制数值做差后的结果以字符串形式输出。
package com.swift; public class String_To_Integer_Test { public static void main(String[] args) { /* ...
- Map<String,Integer>acount字符串出现的次数
随机推荐
- Redis(四)-配置
Redis 配置 Redis 的配置文件位于 Redis 安装目录下,文件名为 redis.conf. 你可以通过 CONFIG 命令查看或设置配置项. 语法 Redis CONFIG 命令格式如下: ...
- python 11：range(起始索引,终止索引,步数)（默认情况下步数为1，生成从起始索引，每次增加(终止索引-起始索引)/步数，到终止索引前的数字串）
squares = [] for value in range(1,11): #第三参数默认为1,生成从1开始,每次增加1步数,到11前的10为止的数字串 square = value ** 2 sq ...
- X - Vasya and Socks
Problem description Vasya has n pairs of socks. In the morning of each day Vasya has to put on a pai ...
- A - Design Tutorial: Learn from Math（哥德巴赫猜想）
Problem description One way to create a task is to learn from math. You can generate some random mat ...
- MySQL定期执行任务相关问题
在sqlyog某数据库下的事件里新建事件,并写入一下代码: DELIMITER $$ ALTER DEFINER=`root`@`%` EVENT `0` ON SCHEDULE EVERY 24 H ...
- HTML链接用法
1.链接的打开方式 ①在新页面打开 <a href="www.baidu.com" target="_blank">百度一下</a> ② ...
- Hadoop学习笔记（一）Hadoop的单节点安装
要想深入学习Hadoop分布式文件系统,首先需要搭建Hadoop的实验环境,Hadoop有两种安装模式,即单节点集群模式安装(也称为伪分布式)和完全分布式模式安装,本节只介绍单节点模式的安装,参考官方 ...
- 关于Core里的 StartUp里的方法的理解。
public class Startup { public Startup(IConfiguration configuration) { Configuration = configuration; ...
- 【sqli-labs】 less11 POST - Error Based - Single quotes- String (基于错误的POST单引号字符型注入)
查看源码,用户名和密码通过post提交 加单引号提交 出现报错,推测对应的SQL语句 , 直接使用or构造永真登录 成功,注意此处登录的用户为表中的第一个用户 需要改变用户可以通过改变筛选条件实现 作 ...
- （转）Arcgis for JS之地图自适应调整
http://blog.csdn.net/gisshixisheng/article/details/42675897 概述:本节讲述的内容为当浏览器大小发生变化或者地图展示区域的大小发生变化时,地图 ...