• 前言
• 数据
• 实现构造无限级
• 广度优先遍历
• 深度优先遍历

前言

最近接到一个关于树的需求，想起了大学那会儿听过俞敏洪有关树的演讲，老师重复在课上放了很多次，N年过去了印象还是很深。
演讲内容是介样的，让我们一起来喝汤~

回到今天的主题，需求是这样的：构造无限级树型结构，界面类似下边这样的结构

然后，还需要把所有的橙色标记的叶子节点从垂直方向筛选出来，最终效果就是得到一个这样顺序的一级数组

KKKQQQJJJHHHGGGLLL

需要根据AAA这个节点和他所有子节点的数据，构造出一个无限级树型结构，如下

数据

• 顶级AAA数据

{"id":348019,"name":"AAA","is_leaf":2,"parent_id":347234,"type":0,"order":["348020"]}

• AAA所有的孩子节点

[{"id":348020,"name":"BBB","is_leaf":2,"parent_id":348019,"type":0,"order":["348037","348033","348021","348023","348034"]},{"id":348021,"name":"CCC","is_leaf":2,"parent_id":348020,"type":0,"order":["348022","348035"]},{"id":348022,"name":"DDD","is_leaf":2,"parent_id":348021,"type":0,"order":["348024","348032","348030"]},{"id":348024,"name":"EEE","is_leaf":2,"parent_id":348022,"type":0,"order":["348025","348036"]},{"id":348025,"name":"FFF","is_leaf":2,"parent_id":348024,"type":0,"order":["348031"]},{"id":348030,"name":"HHH","is_leaf":1,"parent_id":348022,"type":11,"order":[]},{"id":348023,"name":"GGG","is_leaf":1,"parent_id":348020,"type":11,"order":[]},{"id":348031,"name":"QQQ","is_leaf":1,"parent_id":348025,"type":11,"order":[]},{"id":348032,"name":"JJJ","is_leaf":1,"parent_id":348022,"type":11,"order":[]},{"id":348035,"name":"MMM\u65e0\u53f6\u5b50","is_leaf":2,"parent_id":348021,"type":0,"order":[]},{"id":348036,"name":"NNN\u65e0\u53f6\u5b50","is_leaf":2,"parent_id":348024,"type":0,"order":[]},{"id":348033,"name":"KKK","is_leaf":1,"parent_id":348020,"type":11,"order":[]},{"id":348037,"name":"OOO","is_leaf":2,"paread8nt_id":348020,"type":0,"order":[]},{"id":348034,"name":"LLL","is_leaf":1,"parent_id":348020,"type":11,"order":[]}]

节点数据库结构都一样，核心是有一个parent_id，每个节点会保存直属叶子的排序order

实现

其实这里考察的递归构造树、广度优先遍历、深度优先遍历
其实不管啥语言，思想都是这样，因为这个需求是php系统写的接口，就用php给出实现代码，气氛搞起来~

构造无限级

private function getTree($list, $node){$tree = [];foreach ($list as $k => $v) {if ($v['parent_id'] == $node['id']) {$v['children'] = $this->getTree($list, $v);$tree[] = $v;}}if (empty($node['order'])) {return [];}//注意：为了保证所有的叶子节点是根据已有的order排序，这里我们重新摆放一下$treeMap = [];foreach ($tree as $v) {$treeMap[$v['id']] = $v;//建立id和数据的map}//根据order重新排序$orderTree = [];foreach ($node['order'] as $id) {//根据order的顺序重新摆放$orderTree[] = $treeMap[$id];}return $orderTree;}

我们来使用一下，使用如下

最终输出的json如下

可以看到，已经构造出了一个无限级的树，并且和我们的界面是一模一样的层级和顺序

拉下来现在我们要遍历这棵树，为了获取到所有is_leaft=1的节点，我们先用广度优先遍历来试一下，因为这个对于找节点来说性能比较高

广度优先遍历

代码如下

    private function bfsFindLeafs($node){$result = [];//满足条件的组合$q = [];$q[] = $node;while (count($q) > 0) {$size = count($q);for ($i = 0; $i < $size; $i++) {$current = array_pop($q);//判断节点如果是满足条件，加入路径if ($current['is_leaf'] == 1) {$result[] = $current;}//把所有的子节点加入队列foreach ($current['children'] as $v) {$q[] = $v;}}}56creturn $result;}

我们来使用一下

$getAllLeafs = $this->bfsFindLeafs($topNode);echo json_encode($getAllLeafs);

运行后输出

[{"id": 348034,"name": "LLL","is_leaf": 1,"parent_id": 348020,"type": 11,"order": [],"children": []},{"id": 348023,"name": "GGG","is_leaf": 1,"parent_id": 348020,"type": 11,"order": [],"children": []},{"id": 348030,"name": "HHH","is_leaf": 1,"parent_id": 348022,"type": 11,"order": [],"children": []},{"id": 348032,"name": "JJJ","is_leaf": 1,"parent_id": 348022,"type": 11,"order": [],"children": []},{"id": 348031,"name": "QQQ","is_leaf": 1,"parent_id": 348025,"type": 11,"order": [],"children": []},{"id": 348033,"name": "KKK","is_leaf": 1,"parent_id": 348020,"type": 11,"order": [],"children": []}]

好，我们已经可以获取到所有is_leaf为1的节点了，但是这里有一个问题，不是从上到下输出的

LLLGGGHHHJJJQQQKKK

深度优先遍历

深度优先遍历其实就是回滚算法的一种，为了从上到下输出，需要先序遍历

    private function dfsFindLeafs($node){$trackList = [];foreach ($node['children'] as $v) {if ($v['is_leaf'] == 1) {$trackList[] = $v;}if (empty($v['children'])) {continue;}$trackList = array_merge($trackList,$this->dfsFindLeafs($v));}return $trackList;}

我们再来运行一下

$getAllLeafs = $this->dfsFindLeafs($topNode);echo json_encode($getAllLeafs);

输出

[{"id": 348033,"name": "KKK","is_leaf": 1,"parent_id": 348020,"type": 11,"order": [],"children": []},{"id": 348031,"name": "QQQ","is_leaf": 1,"parent_id": 348025,"type": 11,"order": [],"children": []},{"id": 348032,"name": "JJJ","is_leaf": 1,"parent_id": 348022,"type": 11,"order": [],ad0"children": []},{"id": 348030,"name": "HHH","is_leaf": 1,"parent_id": 348022,"type": 11,"order": [],"children": []},{"id": 348023,"name": "GGG","is_leaf": 1,"parent_id": 348020,"type": 11,"order": [],"children": []},{"id": 348034,"name": "LLL","is_leaf": 1,"parent_id": 348020,"type": 11,"order": [],"children": []}]

顺序为

KKKQQQJJJHHHGGGLLL

这次顺序终于对了，oh yeah~，和我们之前的界面垂直顺序一致

完整代码

下面给出完整代码

<?php/*** Class Test* @author chenqionghe*/class Test{public function run(){$topNodeJson = '{"id":348019,"name":"AAA","is_leaf":2,"parent_id":347234,"type":0,"order":["348020"]}';$childrenJson = '[{"id":348020,"name":"BBB","is_leaf":2,"parent_id":348019,"type":0,"order":["348037","348033","348021","348023","348034"]},{"id":348021,"name":"CCC","is_leaf":2,"parent_id":348020,"type":0,"order":["348022","348035"]},{"id":348022,"name":"DDD","is_leaf":2,"parent_id":348021,"type":0,"order":["348024","348032","348030"]},{"id":348024,"name":"EEE","is_leaf":2,"parent_id":348022,"type":0,"order":["348025","348036"]},{"id":348025,"name":"FFF","is_leaf":2,"parent_id":348024,"type":0,"order":["348031"]},{"id":348030,"name":"HHH","is_leaf":1,"parent_id":348022,"type":11,"order":[]},{"id":348023,"name":"GGG","is_leaf":1,"parent_id":348020,"type":11,"order":[]},{"id":348031,"name":"QQQ","is_leaf":1,"parent_id":348025,"type":11,"order":[]},{"id":348032,"name":"JJJ","is_leaf":1,"parent_id":348022,"type":11,"order":[]},{"id":348035,"name":"MMM\u65e0\u53f6\u5b50","is_leaf":2,"parent_id":348021,"type":0,"order":[]},{"id":348036,"name":"NNN\u65e0\u53f6\u5b50","is_leaf":2,"parent_id":348024,"type":0,"order":[]},{"id":348033,"name":"KKK","is_leaf":1,"parent_id":348020,"type":11,"order":[]},{"id":348037,"name":"OOO","is_leaf":2,"parent_id":348020,"type":0,"order":[]},{"id":348034,"name":"LLL","is_leaf":1,"parent_id":348020,"type":11,"order":[]}]';$topNode = json_decode($topNodeJson, true);$childrenList = json_decode($childrenJson, true);$topNode['children'] = $this->getTree($childrenList, $topNode);$getAllLeafs = $this->dfsFindLeafs($topNode);echo json_encode($getAllLeafs);}/*** 递归构造无限级树** @param $li19d6st* @param $node* @return array*/private function getTree($list, $node){$tree = [];foreach ($list as $k => $v) {if ($v['parent_id'] == $node['id']) {$v['children'] = $this->getTree($list, $v);$tree[] = $v;}}if (empty($node['order'])) {return [];}//注意：为了保证所有的叶子节点是根据已有的order排序，这里我们重新摆放一下$treeMap = [];foreach ($tree as $v) {$treeMap[$v['id']] = $v;//建立id和数据的map}//根据order重新排序$orderTree = [];foreach ($node['order'] as $id) {//根据order的顺序重新摆放$orderTree[] = $treeMap[$id];}return $orderTree;}/*** 广度优先遍历获取所有叶子* @param $node* @return array*/private function bfsFindLeafs($node){$result = [];//满足条件的组合$q = [];$q[] = $node;while (count($q) > 0) {$size = count($q);for ($i = 0; $i < $size; $i++) {$current = array_pop($q);//判断节点如果是满足条件，加入路径if ($current['is_leaf'] == 1) {$result[] = $current;}//把所有的子节点加入队列foreach ($current['children'] as $v) {$q[] = $v;}}}return $result;}/*** 获取所有的叶子路径（深度度优先遍历）*/private function dfsFindLeafs($node){$trackList = [];foreach ($node['children'] as $v) {if ($v['is_leaf'] == 1) {$trackList[] = $v;}if (empty($v['children'])) {continue;}$trackList = array_merge($trackList, $this->dfsFindLeafs($v));}return $trackList;}}(new Test())->run();

ok，到这里我们就已经学会如何构造无限级树型结构，并学会用深度优先遍历垂直输出指定条件节点，还知道了怎么用广度优先遍历，giao~