生成树的一般算法为遍历集合,获取父类,再依次递归遍历集合获取其子类,这次的升级算法大致如此,但是在遍历之前先进行了一次分组,将所有父级id相同的分为一组,遍历父级时直接根据id从map中获取其子级,这样一来,就之遍历了n次,算上分组遍历的n次,时间复杂度仅为O(2n);
树工厂类:
/**
* 树工厂类,用于生成树
* @author xieshuang
* @date 2019-08-08 10:43
*/
public class TreeFactory{
int a;
/**
* 优化方案
* 1.按照pid进行分组
* @param treeNodes
* @return
*/
public CollectioncreateTree(CollectiontreeNodes){
a = 0;
long l2 = System.currentTimeMillis();
Map> collect = treeNodes.stream().collect(Collectors.groupingBy(T::getPid));
System.out.println("第一次遍历花费:"+(System.currentTimeMillis()-l2));
long l3 = System.currentTimeMillis();
CollectiontreeNodeList = new ArrayList<>();
Listlist;
Listlist1 = collect.get(0);
Listlist2 = collect.get(0L);
if (list1==null){
list = list2;
}else {
list = list1;
}
if (list!=null){
for (T t : list) {
treeNodeList.add(addChildNode(t, collect));
}
}
System.out.println("第二次遍历花费:"+(System.currentTimeMillis()-l3));
System.out.println("总共遍历了:"+a);
return treeNodeList;
}
private T addChildNode(T treeNode, Map> collect){
a++;
Listlist = collect.get(treeNode.getId());
if (list!=null){
ListtreeNodeList = new ArrayList<>();
for (T t : list) {
treeNodeList.add(addChildNode(t, collect));
}
treeNode.setChilds(treeNodeList);
}
return treeNode;
}
}
树父类:
@Data
public class Tree{
private Number id;
private Number pid;
private Collection childs;
}
自定义的树类:
@Data
public class MyTree extends Tree {
private String name;
}
public class TreeFactoryTest {
public static void main(String[] args) {
Listtrees = new ArrayList<>();
for (int i = 1; i < 100; i++) {
MyTree myTree = new MyTree();
myTree.setName("顶级"+i);
myTree.setId(i);
myTree.setPid(0);
trees.add(myTree);
for (int j = 0; j <100 ; j++) {
MyTree myTree1 = new MyTree();
myTree1.setName("子级"+j);
myTree1.setId(j+(i*100));
myTree1.setPid(i);
trees.add(myTree1);
for (int k = 0; k <100 ; k++) {
MyTree myTree2 = new MyTree();
myTree2.setName("子级"+j);
myTree2.setId(k+(j*100)+(i*10000));
myTree2.setPid(j+(i*100));
trees.add(myTree2);
}
}
}
System.out.println(trees.size());
Collectiontree = new TreeFactory().createTree(trees);
//System.out.println(JSONObject.toJSONString(tree));
}
}
测试结果:
百万数据量耗时0.2秒