获取最新citysql 城市sql 国家统计局最新市区分布 thinkphp php 抓取

    /**
     * # +========================================================================
     * # | - @name        抓取全国统计用区划代码和城乡划分代码
     * # | - @author     cq <just_leaf@foxmail.com> 
     * # | - @copyright zmtek 2019-12-26
     * # +------------------------------------------------------------------------
     * # | - 1.http://www.stats.gov.cn/ - 国家统计局官网
     * # +========================================================================
     */
    public function getcity() {
        
        header("Content-type: text/html; charset=gb2312");
        # 超时设置
        set_time_limit(0);

        $index = file_get_contents("http://www.stats.gov.cn/tjsj/tjbz/tjyqhdmhcxhfdm/2019/index.html");

        $url = 'http://www.stats.gov.cn/tjsj/tjbz/tjyqhdmhcxhfdm/2019/';
        error_reporting(0);
        $prov = array(
            array(),
            array(11, 12, 13, 14, 15, 21, 22, 23, 31, 32, 33, 34, 35, 36, 37, 41, 42, 43, 44, 45, 46, 50, 51, 52, 53, 54, 61, 62, 63, 64, 65),
            array('北京市', '天津市', '河北省', '山西省', '内蒙古自治区', '辽宁省', '吉林省', '黑龙江省',
                '上海市', '江苏省', '浙江省', '安徽省', '福建省', '江西省', '山东省', '河南省',
                '湖北省', '湖南省', '广东省', '广西壮族自治区', '海南省', '重庆市', '四川省', '贵州省',
                '云南省', '*自治区', '陕西省', '甘肃省', '青海省', '宁夏回族自治区', '**自治区',
            ),
        );

        $matches = $prov;
        $i = 0; 
        for ($i = 0, $e = count($matches[1]); $i < $e; $i++)
        {
            
            preg_match_all('/<a href=\'\d{2}\/(.{1,30}).html\'>(.{1,30})<\/a><\/td><\/tr>/', $index, $matche);
            for ($a = 0, $b = count($matche[1]); $a < $b; $a++)
            {    

                $index = file_get_contents($url . $matches[1][$i] . '/' . $matche[1][$a] . '.html');
                preg_match_all('/<a href=\'\d{2}\/(.{1,30}).html\'>(.{1,30})<\/a><\/td><\/tr>/', $index, $match);
                for ($c = 0, $d = count($match[1]); $c < $d; $c++)
                {
                    //补0处理
                    $provinces_id      = str_pad($matches[1][$i], 12, "0", STR_PAD_RIGHT);
                    $province_name     = iconv("GB2312", "UTF-8", $matches[2][$i]);
                    $city_id         = str_pad($matche[1][$a], 12, "0", STR_PAD_RIGHT);
                    $city_name         = iconv("GB2312", "UTF-8", $matche[2][$a]);
                    
                    $aru   = substr($matche[1][$a], 2, 2);
                    $index = file_get_contents($url . $matches[1][$i] . '/' . $aru . '/' . $match[1][$c] . '.html');
                    preg_match_all('/<a href=\'\d{2}\/(.{1,30}).html\'>(.{1,30})<\/a><\/td><\/tr>/', $index, $matc);
                    //部分省市的html和大部分的不一样,重写规则
                    if(!$matc[0]) preg_match_all('/<td>(.{1,30})<\/td><td>\d{1,10}<\/td><td>(.{1,30})<\/td><\/tr>/', $index, $matc);
                    for ($v = 0, $n = count($matc[1]); $v < $n; $v++)
                    {
                        $county_id     = str_pad($match[1][$c], 12, "0", STR_PAD_RIGHT);
                        $county_name= iconv("GB2312", "UTF-8", $match[2][$c]);
                        $town_id     = str_pad($matc[1][$v], 12, "0", STR_PAD_RIGHT);
                        $town_name     = iconv("GB2312", "UTF-8", $matc[2][$v]);
                        
                        $aru2   = substr($matche[1][$a],-2);
                        $index = file_get_contents($url . $matches[1][$i] . '/' . $aru . '/' .$aru2 .'/'. $matc[1][$c] . '.html');
                        preg_match_all('/<a href=\'\d{2}\/(.{1,30}).html\'>(.{1,30})<\/a><\/td><\/tr>/', $index, $mat);
                        //部分省市的html和大部分的不一样,重写规则
                        if(!$mat[0]) preg_match_all('/<td>(.{1,30})<\/td><td>\d{1,10}<\/td><td>(.{1,30})<\/td><\/tr>/', $index, $mat);
                        
                        if(!$mat[0]) {

                            $add = array(
                                'province_id'     => $provinces_id ,
                                'province_name'    => $province_name ,
                                'city_id'         => $city_id ,
                                'city_name'     => $city_name ,
                                'county_id'     => $county_id ,
                                'county_name'     => $county_name ,
                                'town_id'         => $town_id,
                                'town_name'        => $town_name 
                            );

                            M('position') -> add($add);
                        }else{
                            for($z = 0, $x = count($mat[1]); $z < $x; $z++){
                                
                                $housing_id     = $mat[1][$z];
                                $housing_name     = iconv("GB2312", "UTF-8", $mat[2][$z]);
                                
                                $add = array(
                                    'province_id'     => $provinces_id ,
                                    'province_name'    => $province_name ,
                                    'city_id'         => $city_id ,
                                    'city_name'     => $city_name ,
                                    'county_id'     => $county_id ,
                                    'county_name'     => $county_name ,
                                    'town_id'         => $town_id,
                                    'town_name'        => $town_name ,
                                    'housing_id'     => $housing_id ,
                                    'housing_name'     => $housing_name ,
                                );

                                M('position') -> add($add);
                            }
                        }
                    }
                }
            }
        }
    }
CREATE TABLE `gd_position` (
  `id` int(10) unsigned NOT NULL AUTO_INCREMENT,
  `province_id` varchar(20) DEFAULT NULL,
  `province_name` varchar(255) CHARACTER SET utf8mb4 DEFAULT '',
  `city_id` varchar(20) DEFAULT NULL,
  `city_name` varchar(255) CHARACTER SET utf8mb4 DEFAULT '',
  `county_id` varchar(20) DEFAULT NULL,
  `county_name` varchar(255) CHARACTER SET utf8mb4 DEFAULT '',
  `town_id` varchar(20) DEFAULT NULL,
  `town_name` varchar(255) CHARACTER SET utf8mb4 DEFAULT '',
  `housing_id` varchar(20) DEFAULT NULL,
  `housing_name` varchar(255) DEFAULT NULL,
  PRIMARY KEY (`id`)
) ENGINE=MyISAM AUTO_INCREMENT=821 DEFAULT CHARSET=utf8;

 

上一篇:《单词的减法》state1~state17(200p)


下一篇:Sorting Non-numeric Data in R & Python