java获取行政区划获取到街道

前言

这几短时间需要获取行政区划，在网上看了很多的网站大部分都是直接网页直接加载，一个一个界面获取太麻烦所以自己录了段代码，自己获取行政区划。
之前试着做了一段java的爬虫，很不幸没做成功，但是这也对我这段代码做了铺垫吧。数据处理都是那时候学的，所以处理起来不算太难。

代码获取数据

引入依赖

引入jsoup用来处理页面数据

     <dependency>
            <groupId>org.jsoup</groupId>
            <artifactId>jsoup</artifactId>
            <version>1.12.1</version>
     </dependency>

构建数据类

构建基础类用来存放

 	@Data
    @Builder
    static class Area {
        /**
         * 街道编码
         */
        private String streetCode;
        /**
         * 街道名称
         */
        private String streetName;
        /**
         * 区县编码
         */
        private String districtCode;
        /**
         * 区县名称
         */
        private String districtName;
        /**
         * 城市编码
         */
        private String cityCode;
        /**
         * 城市名称
         */
        private String cityName;
        /**
         * 省份编码
         */
        private String provinceCode;
        /**
         * 省份名称
         */
        private String provinceName;
    }

获取全部省份数据

从这里获取全部省份的数据

这个界面中全部数据主要获取的每个省的行政代码，因为行政代码需要用来获取每个省的数据的相关参数。从金额面代码中可以看出来，根据id为’example‘的节点，可以获取数据的表格，然后根据标签获取每一行数据，然后根据td获取每个节点的数据

private List<Area> getData() {
        try {
            URL url = new URL("https://www.tianzhicun.com");
            HttpsURLConnection con = (HttpsURLConnection) url.openConnection();
            // https请求
            X509TrustManager xtm = new X509TrustManager() {
                @Override
                public X509Certificate[] getAcceptedIssuers() {
                    // TODO Auto-generated method stub
                    return null;
                }
                @Override
                public void checkServerTrusted(X509Certificate[] arg0, String arg1)
                        throws CertificateException {
                    // TODO Auto-generated method stub
                }
                @Override
                public void checkClientTrusted(X509Certificate[] arg0, String arg1)
                        throws CertificateException {
                    // TODO Auto-generated method stub
                }
            };
            TrustManager[] tm = {xtm};
            SSLContext ctx = SSLContext.getInstance("TLS");
            ctx.init(null, tm, null);

            con.setSSLSocketFactory(ctx.getSocketFactory());
            con.setHostnameVerifier(new HostnameVerifier() {
                @Override
                public boolean verify(String arg0, SSLSession arg1) {
                    return true;
                }
            });
            
            InputStream inStream = con.getInputStream();
            ByteArrayOutputStream outStream = new ByteArrayOutputStream();
            byte[] buffer = new byte[1024];
            int len = 0;
            while ((len = inStream.read(buffer)) != -1) {
                outStream.write(buffer, 0, len);
            }
            byte[] b = outStream.toByteArray();//网页的二进制数据
            outStream.close();
            inStream.close();
            // 返回的html界面字符串
            String rtn = new String(b, "utf-8");
            Document document = Jsoup.parse(rtn);
            // 获取example下的表格
            Element postList = document.getElementById("example");
            // 获取tr元素每一行数据代表一条数据
            Elements postItems = postList.getElementsByTag("tr");
            //存储查询到的数据
            List<Area> list = new ArrayList<>();
            // 前两个数据为标题头的数据所以舍弃掉了
            int i = 0;
            for (Element postItem : postItems) {
                i++;
                if (i < 3) {
                    continue;
                }
                Area area = Area.builder()
                        .provinceCode(postItem.getElementsByTag("td").get(1).text())
                        .build();
                list.add(area);
            }
            return list;
        } catch (Exception e) {
            e.printStackTrace();
            System.out.println(string);
        }
        return null;
    }

获取每个省份的数据

  List<Area> list = new ArrayList<>();
  for (Area area : areas) {
            String code = area.getProvinceCode();
			// 处理逻辑和上面相同
             URL url = new URL("https://www.tianzhicun.com/province/" + code + "/town");
            HttpsURLConnection con = (HttpsURLConnection) url.openConnection();
            X509TrustManager xtm = new X509TrustManager() {
                @Override
                public X509Certificate[] getAcceptedIssuers() {
                    // TODO Auto-generated method stub
                    return null;
                }

                @Override
                public void checkServerTrusted(X509Certificate[] arg0, String arg1)
                        throws CertificateException {
                    // TODO Auto-generated method stub

                }

                @Override
                public void checkClientTrusted(X509Certificate[] arg0, String arg1)
                        throws CertificateException {
                    // TODO Auto-generated method stub

                }
            };

            TrustManager[] tm = {xtm};

            SSLContext ctx = SSLContext.getInstance("TLS");
            ctx.init(null, tm, null);

            con.setSSLSocketFactory(ctx.getSocketFactory());
            con.setHostnameVerifier(new HostnameVerifier() {
                @Override
                public boolean verify(String arg0, SSLSession arg1) {
                    return true;
                }
            });


            InputStream inStream = con.getInputStream();
            ByteArrayOutputStream outStream = new ByteArrayOutputStream();
            byte[] buffer = new byte[1024];
            int len = 0;
            while ((len = inStream.read(buffer)) != -1) {
                outStream.write(buffer, 0, len);
            }
            byte[] b = outStream.toByteArray();//网页的二进制数据
            outStream.close();
            inStream.close();
            String rtn = new String(b, "utf-8");
//            System.out.println(rtn);
            Document document = Jsoup.parse(rtn);
            //像js一样，通过标签获取title
//            System.out.println(document.getElementsByTag("title").first());
            //像js一样，通过id 获取文章列表元素对象
            Element postList = document.getElementById("example");
            //像js一样，通过class 获取列表下的所有博客
            Elements postItems = postList.getElementsByTag("tr");
       
            int i = 0;
            for (Element postItem : postItems) {
                i++;
                if (i < 3) {
                    continue;
                }
              //处街道编码，区域编码都在<a>标签中，并且带有链接 需哦一需要对器做特殊处理 先获取a标签在获取线面href的链接在把非编码替换掉
                Area area = Area.builder()
                        .streetName(postItem.getElementsByTag("td").get(0).text())
                        .streetCode(postItem.getElementsByTag("td").get(1).text())                        
                        .districtCode(postItem.getElementsByTag("td").get(5).getElementsByTag("a").attr("href").replace("/province/city/county/", ""))
                        .districtName(postItem.getElementsByTag("td").get(5).text())
                        .cityCode(postItem.getElementsByTag("td").get(4).getElementsByTag("a").attr("href").replace("/province/city/", ""))
                        .cityName(postItem.getElementsByTag("td").get(4).text())
                        .provinceCode(postItem.getElementsByTag("td").get(3).getElementsByTag("a").attr("href").replace("/province/", ""))
                        .provinceName(postItem.getElementsByTag("td").get(3).text())
                        .build();
                list.add(area);
            }
          
        } catch (Exception e) {
            e.printStackTrace();
            System.out.println(string);
        }
      }

此时数据已经获取完毕
此时json很大所以添加了保存到文件中

文件保存

使用poi展示数据

 		HSSFWorkbook wb = new HSSFWorkbook();
        // 全部区域到街道
        HSSFSheet streetSheet = wb.createSheet("街道");
        for (int i = 0; i < list.size(); i++) {
            Area area = list.get(i);
            HSSFRow streetRow = streetSheet.createRow(i);
            HSSFCell cell00 = streetRow.createCell(0);
            cell00.setCellValue(area.getProvinceName());
            HSSFCell cell10 = streetRow.createCell(1);
            cell10.setCellValue(area.getProvinceCode());
            HSSFCell cell20 = streetRow.createCell(2);
            cell20.setCellValue(area.getCityName());
            HSSFCell cell30 = streetRow.createCell(3);
            cell30.setCellValue(area.getCityCode());
            HSSFCell cell40 = streetRow.createCell(4);
            cell40.setCellValue(area.getDistrictName());
            HSSFCell cell50 = streetRow.createCell(5);
            cell50.setCellValue(area.getDistrictCode());
            HSSFCell cell60 = streetRow.createCell(6);
            cell60.setCellValue(area.getStreetName());
            HSSFCell cell70 = streetRow.createCell(7);
            cell70.setCellValue(area.getStreetCode());
        }
        FileOutputStream fout = new FileOutputStream("D:/区域.xls");
        wb.write(fout);

在这里插入图片描述