官网: http://trove.starlight-systems.com/overview
trove相比jdk原生的集合类有三个优势:
1、更高的性能
2、更底的内存消耗
3、除了实现原生Collections API并额外提供更强大的功能
集合是java编程最常用的API之一,把项目的集合对象改用trove替换就能获得性能提升和内存的节省。
这是对一个项目底成本的优化方案。
下面二段代码分别测试trove和原生的集合的性能
1. trove
public void testIterator(){ TIntObjectMap map = new TIntObjectHashMap(); for( int i = 0; i < 1000; i++){ Game g = new Game(); g.setName( "最终幻想" + i); g.setSize(15000 + (i << 3)); g.setCtDate( new Date()); map.put(i, g); } int size = map.size(); TIntObjectIterator it = map.iterator(); for( int i = size; i > 0; i--){ it.advance(); System. out.println(it.key() + "=" + it.value()); if(it.key() == 3){ Game g = new Game(); g.setName( "最终幻想13"); g.setSize(15000 + (i << 3)); g.setCtDate( new Date()); it.setValue(g); } } System. out.println( "=======================================" ); System. out.println(map.get(3)); }
2. HashMap
public void testIterator(){ Map map = new HashMap(); for( int i = 0; i < 1000; i++){ Game g = new Game(); g.setName( "最终幻想" + i); g.setSize(15000 + (i << 3)); g.setCtDate( new Date()); map.put(i, g); } Set set = map.entrySet(); for( Iterator it = set.iterator(); it.hasNext(); ){ Entry e = (Entry )it.next(); System. out.println(e.getKey() + "=" + e.getValue()); if(((Integer)e.getKey()).intValue() == 3){ Game g = new Game(); g.setName( "最终幻想13"); g.setSize(18000); g.setCtDate( new Date()); e.setValue(g); } } System. out.println( "=======================================" ); System. out.println(map.get(3)); }
两段的测试结果,trove花费0.001s,HashMap花费0.090s
trove的测试代码使用其Map额外提供的iterator进行遍历,当然他也支持通过entrySet方式遍历。
使用SizeOf.jar计算上面两段的map占用内存的大小,trove占用214024byte,HashMap占用250128byte,可见内存方面节省的比率并是很大。
从官方下载的API中可以看中trove根据基础的数据类型实现了Map、List和Set,比如TIntObjectHashMap其key只能是int, value是继承Object的所有对象;TIntList只能存入int;TIntSet只能存入int。
在java里有些对象的hashCode和equals方法不能被覆盖(比如:String和数组),以这种对象为Map的key时就不能随心所欲,trove的提供的custom map可以解除这种约束,只需要实现HashingStrategy接口传递给Map,Map判断key是否相同时调用HashingStrategy的computeHashCode和equals代替object自身的hashCode和equals,以下是官网一个例子
For example, this code:
char[] foo, bar; foo = new char[] {'a','b','c'}; bar = new char[] {'a','b','c'}; System.out.println(foo.hashCode() == bar.hashCode() ? "equal" : "not equal"); System.out.println(foo.equals(bar) ? "equal" : "not equal");
produces this output:
not equal
not equal
And so an entry stored in a java.util.HashMap with foo as a key could not be retrieved with bar, since there is no way to override hashCode() or equals() on language array objects.
In a gnu.trove.THashMap, however, you can implement a TObjectHashingStrategy to enable hashing on arrays:
class CharArrayStrategy implements TObjectHashingStrategy { public int computeHashCode(Object o) { char[] c = (char[])o; // use the shift-add-xor class of string hashing functions // cf. Ramakrishna and Zobel, "Performance in Practice // of String Hashing Functions" int h = 31; // seed chosen at random for (int i = 0; i < c.length; i++) { // could skip invariants h = h ^ ((h << 5) + (h >> 2) + c[i]); // L=5, R=2 works well for ASCII input } return h; } public boolean equals(Object o1, Object o2) { char[] c1 = (char[])o1; char[] c2 = (char[])o2; if (c1.length != c2.length) { // could drop this check for fixed-length keys return false; } for (int i = 0, len = c1.length; i < len; i++) { // could skip invariants if (c1[i] != c2[i]) { return false; } } return true; } }
总结:
Trove是很值得了解和使用的一个开源项目,你会喜欢上他的。