php 抓取中国统计局 最新县及县以上行政区划代码

起因：

　　前两天突然想找个省市县的行政代码库，发现网上要么不是最新的，要么要帐号，要积分，要钱。让人好烦，就写了这个脚本。

数据库结构：

CREATE TABLE IF NOT EXISTS `area` (
  `id` int(11) NOT NULL auto_increment,
  `code` varchar(6) NOT NULL,
  `name` varchar(20) NOT NULL,
  `citycode` varchar(6) NOT NULL,
  PRIMARY KEY  (`id`)
) ENGINE=Innodb  DEFAULT CHARSET=utf8;
 
CREATE TABLE IF NOT EXISTS `city` (
  `id` int(11) NOT NULL auto_increment,
  `code` varchar(6) NOT NULL,
  `name` varchar(20) NOT NULL,
  `provincecode` varchar(6) NOT NULL,
  PRIMARY KEY  (`id`)
) ENGINE=Innodb  DEFAULT CHARSET=utf8;

 
CREATE TABLE IF NOT EXISTS `province` (
  `id` int(11) NOT NULL auto_increment,
  `code` varchar(6) NOT NULL,
  `name` varchar(20) NOT NULL,
  PRIMARY KEY  (`id`)
) ENGINE=Innodb  DEFAULT CHARSET=utf8 ;

脚本文件：

<?php
set_time_limit(0);
/**
 * 
 */
class get_city_code {
    //
    private $html = '';
    public  $code_rt;
    private static $instance = '';
    private $db = '';
    private $box = array();
    private $url = 'http://www.stats.gov.cn/tjsj/tjbz/xzqhdm/201504/t20150415_712722.html';
    //
    private function __construct() {
        
    }

    public static function getInstance() {
        if ( ! self::$instance instanceof get_city_code ) {
            self::$instance = new self();
        }
        return self::$instance;
    }
    public  function start() {
        //
        $this->connect_tongji_html();
        $this->code_rt = new code_result();
        $this->code_rt->html = $this->html;
        $this->code_rt->filter_all_data();
    }
    private function connect_tongji_html() {
        $ch = curl_init();
        $url = $this->getUrl();

        curl_setopt($ch, CURLOPT_URL, $url);
        curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
        curl_setopt($ch, CURLOPT_HEADER, 0);
        //执行并获取HTML文档内容
        $this->html = curl_exec($ch);

        //释放curl句柄
        curl_close($ch);
    }

    //获取url
    private function getUrl() {
        return $this->url;
    }

    //插入数据库
    public function insert() {
        $i=0;
        $box = array();
        foreach($this->code_rt->code_data as $k => $v) {
            $k_arr = str_split($k,2);
            $first = $k_arr[0];
            $second = $k_arr[1];
            $three = $k_arr[2];

            if( !empty( $box[$first]) ){
                if( !empty($box[$first][$second]) ) {
                    $box[$first][$second][$three] = $v;
                }else{                    
                    $box[$first][$second][] = $v;
                }
            }else{
                $box[$first][] = array($v);
            }
        }
        
        $this->get_db();
        //print_r($box);
        foreach( $box as $k1=>$v1){

            $code1 = $k1."0000";
            $name1 = $v1[0][0];
            $sql = "insert into province values (NULL,'".$code1."','".$name1."')";
            $this->db->query($sql);
            foreach ( $v1 as $k2 => $v2 ) {
                if($k2 == 0) {
                    continue;
                }
                $code2 = $k1.$k2."00";
                $name2 = $v2[0]=='市辖区' ? $name1 : $v2[0];
                if( $name2 == '县'){
                    continue;
                }
                $sql = "insert into city values (NULL,'".$code2."','".$name2."','".$code1."')";
                $this->db->query($sql);

                foreach( $v2 as $k3=>$v3 ) {
                    if($k3 == 0){
                        continue;
                    }
                    $code3 = $k1.$k2.$k3;
                    $name3 = $v3;
                    $sql = "insert into area values (NULL,'".$code3."','".$name3."','".$code2."')";
                    $this->db->query($sql);
                }
            }
        }

        $this->db->close();
    }

    //
    private function get_db () {
        $db = new mysqli('localhost','root','sunl','blog');
        $db->set_charset('utf8');
        $this->db = $db;
    }
}

class code_result {
    public $html = '';
    public $code_data = array();
    private $code_arr  = array();
    private $name_arr  = array();

    public function __construct () {

    }

    public function filter_all_data() {
        //获取所有的p标签
        $patten = "/<p.*><\/p>/";
        preg_match($patten, $this->html, $p);
        //去除所有的&nbsp;
        $this->html = preg_replace( '/&nbsp;/', '', strip_tags($p[0]));

        $this->html = preg_replace( '/\s+/', '', $this->html );
        $this->html = preg_replace( '/　/', '', $this->html );
        $this->html = preg_replace( '/　/', '', $this->html );
        $this->html = preg_replace( '/　/', '', $this->html );
        //echo $this->html;
        
        //匹配code id
        $patten2 = "/[\d{6}]+/";
        preg_match_all($patten2, $this->html, $this->code_arr);
        
        //匹配县市名称
        $patten3 = "/[\x{4e00}-\x{9fa5}]+/u";
        preg_match_all($patten3, $this->html, $this->name_arr);
        
        $this->code_data = array_combine($this->code_arr[0] , $this->name_arr[0]);
        //print_r($this->code_data);die;
    }

    public function getCodeData() {
        return $this->code_data;
    }
}

$code = get_city_code::getInstance();
$code->start();
$code->insert();
?>