3数据分析每日进度报告01

题目要求：

基本完成地域维度标准化，将地址更新为省市县三级，并将行政区划代码添加。

效果图：

关键功能代码片段：

public static String addressResolution(String address){
        String regex="(?<province>[^省]+自治区|.*?省|.*?行政区|.*?市)(?<city>[^市]+自治州|.*?地区|.*?行政单位|.+盟|市辖区|.*?市|.*?县|.+区|.+市)(?<county>[^县]+县|.+区|.+市|.+旗|.+海域|.+岛)";
        Matcher m=Pattern.compile(regex).matcher(address);
        String province=null,city=null,county=null;
        String str="";
        while(m.find()){
            province=m.group("province");
            city=m.group("city");
            county=m.group("county");
            str = province + city + county;
            System.out.println(str);
        }
        return str;
    }

public static String getlocation1(String loc){
        String location2="";
          String url="http://api.map.baidu.com/geocoding/v3/?address="+loc+"10号&output=xml&ak=ld0uqubVfSTAUlXH5qIMN2F3Snsp16LU&callback=showLocation";
          System.out.println(url);
          Document doc = null;
           java.net.HttpURLConnection conn = null;
           InputStream ins = null;
           SAXReader reader = null;
           try{
            //HttpTimeoutHandler hth = new HttpTimeoutHandler(600000);
            URL conURL = new URL(null,url);
            conn = (HttpURLConnection)conURL.openConnection();
            conn.setDoInput(true);
            conn.setDoOutput(true);
            conn.setUseCaches(false);
            ins = conn.getInputStream();
            reader =new SAXReader();
            doc= reader.read(ins);
            //System.out.println(url);
            Element root=doc.getRootElement();
            String docXmlText=doc.asXML();
            //System.out.println(docXmlText);
            Element e=root.element("result");
            Element location=e.element("location");
            Element lng=location.element("lng");
            Element lat=location.element("lat");
            String lng1=lng.asXML();
            String lat1=lat.asXML();
            System.out.println("lng"+lng1);
            System.out.println("lat"+lat1);
           // System.out.println("location"+location.asXML());
            //System.out.println("xiayukun"+e.asXML());
            lng1=lng1.substring(lng1.indexOf("<lng>")+5,lng1.indexOf("</lng>"));
            
            System.out.println(lng1);
            lat1=lat1.substring(lat1.indexOf("<lat>")+5,lat1.indexOf("</lat>"));
            System.out.println(lat1);
            location2=getLocation(lat1,lng1);
            List<Element> list = root.elements("location");
            System.out.println(url);
            for (Element object : list) {
                System.out.println(url);
                System.out.println(object.getName());
                for (Element element : (List<Element>) object.elements()) {
                    System.out.print(((Element) element).getName() + ":");
                    System.out.print(element.getText() + " ");
                }
                System.out.println();

            }

            ins.close();
            conn.disconnect();
           }catch (MalformedURLException e) {
            e.printStackTrace();
           } catch (IOException e) {
            e.printStackTrace();   
           } catch (DocumentException e) {
            e.printStackTrace();
           }catch(Exception e){
            e.printStackTrace();
           }finally {
            try {
             if (ins != null) {
              ins.close();
              ins = null;
             }
            } catch (IOException e1) {
             e1.printStackTrace();
            }
            try {
             if (conn != null) {
              conn.disconnect();
              conn = null;
             }
            } catch (Exception e2) {
             e2.printStackTrace();
            }
           }
           return location2;
        }

 部分运行结果截图：