背景：
公司的行政区划代码有问题，有的没有街道信息，有的关联信息有误，然后找到了国家的网站国家统计局-行政区划，这个里面是包含了所有的行政信息，但是全是html页面，这个就没法自动获取了，只能去爬取这个数据了，java语音有三方类库Jsoup，他是一个仿浏览器的三方库，可以通过他来获取页面信息。
一、导入jar包
下面是笔者用到的全部jar包
<dependency>
<groupId>org.apache.poi</groupId>
<artifactId>poi-ooxml</artifactId>
<version>3.9</version>
</dependency>
<dependency>
<groupId>com.google.guava</groupId>
<artifactId>guava</artifactId>
<version>30.1.1-jre</version>
</dependency>
<dependency>
<groupId>cn.hutool</groupId>
<artifactId>hutool-json</artifactId>
<version>5.4.0</version>
</dependency>
<dependency>
<groupId>com.alibaba</groupId>
<artifactId>fastjson</artifactId>
<version>1.2.44</version>
</dependency>
<dependency>
<groupId>org.jsoup</groupId>
<artifactId>jsoup</artifactId>
<version>1.14.3</version>
</dependency>
<dependency>
<groupId>org.apache.httpcomponents</groupId>
<artifactId>httpclient</artifactId>
<version>4.5.5</version>
</dependency>
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-web</artifactId>
<version>2.5.4</version>
</dependency>
这里说下，maven配置要从阿里云下载jar，若是从中央仓库将会非常的慢。
阿里云私服：http://maven.aliyun.com/nexus/content/repositories/central/
二、代码展示
这里是代码的展示，笔者是网上搜的代码改造的，不然网站有反爬，大概爬取2000条左右就会中断，笔者加了延时这样就避开了反爬（可能还有别的规避措施）。这里爬取的是4级行政区划：省、市、区县、街道
package com.cheng.controller;
import org.apache.poi.ss.usermodel.Row;
import org.apache.poi.xssf.streaming.SXSSFSheet;
import org.apache.poi.xssf.streaming.SXSSFWorkbook;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.select.Elements;
import java.io.FileOutputStream;
import java.io.IOException;
import java.net.ConnectException;
import java.net.SocketTimeoutException;
import java.util.*;
/**
* @author pcc
* @version 1.0.0
* @className JsoupTest
* @date 2023-03-02 10:39
*/
public class JsoupTestPluMdm {
static int i = 1;
static String url1 = "http://www.stats.gov.cn/tjsj/tjbz/tjyqhdmhcxhfdm/2022";
static String url2 = "";
public static void main(String[] args) throws IOException {
try{
List<Map<String,String>> listMap = new ArrayList<>();
Document document = Jsoup.connect(url1)
.header("Cookie", "wzws_sessionid=oGQAAyWBMmNlMWZkgjdlZDJkMIAyMjEuMjM4LjEzMi41MA==; SF_cookie_1=15502425; wzws_cid=6e8cdc0aea81349b05c8a0b6c05cd7204b6e0f10e5a48d462175473d23abcb4891edf1ceb73464398cb1ce7e6f53999f7545dd0014a15b1fb4eec5c6cf37421f0c2b08528de36f728ec4c676ed264c7d")
.get();
//获取他所有的省
Elements elements = document.select("body > table:nth-child(3) > tbody > tr:nth-child(1) > td > table > tbody > tr:nth-child(2) > td > table > tbody > tr > td > table > tbody");
//解析省的超链接
Elements elements1 = elements.select("tbody > tr > td > a");
for(int j=0;j<elements1.size();j++){
// Thread.sleep(100);
String s = elements1.get(j).select("a").attr("href");
String provinceCode = s.replaceAll(".html","")+"0000";
System.out.println("省代码:" + provinceCode);
String provinceName = elements1.get(j).text();
System.out.println("省名称:" + provinceName);
Map<String,String> map = new HashMap<>();
map.put(provinceCode,provinceName);
listMap.add(map);
}
for (int i1 = 0; i1 <31; i1++) {
System.out.println("**********************i********************:"+i);
if(i%1000==0){
Thread.sleep(1000*60*10);
}
Map<String, String> stringStringMap = listMap.get(i1);
Iterator<Map.Entry<String, String>> iterator = stringStringMap.entrySet().iterator();
while(iterator.hasNext()){
Map.Entry<String,String> entry = iterator.next();
String provinceCode = entry.getKey();
String provinceName = entry.getValue();
String index = provinceCode.substring(0,2)+".html";
SXSSFWorkbook wb = new SXSSFWorkbook(100);
SXSSFSheet sheet = (SXSSFSheet) wb.createSheet();
// TODO 这里改成自己的地址即可，也可以存放到一个文件里
String enterFileName = "C:\\Users\\pcc\\Desktop\\xingzhengquhua\\"+provinceName+".xlsx";
FileOutputStream fileOut = new FileOutputStream(enterFileName);
Row row = sheet.createRow(0);
sheet.createRow(i).createCell(0).setCellValue(provinceCode);// id
sheet.getRow(i).createCell(1).setCellValue(provinceName);// name
sheet.getRow(i).createCell(2).setCellValue(""); // pid
sheet.getRow(i).createCell(3).setCellValue("1"); // type
i++;
try {
jsoupList2(url1 + "/" + index, provinceName, provinceCode, sheet);
} catch (SocketTimeoutException e) {
e.printStackTrace();
jsoupList2(url1 + "/" + index, provinceName, provinceCode, sheet);
} catch (ConnectException e) {
e.printStackTrace();
jsoupList2(url1 + "/" + index, provinceName, provinceCode, sheet);
}
row.createCell(0).setCellValue("id");
row.createCell(1).setCellValue("district_name");
row.createCell(2).setCellValue("pid");
row.createCell(3).setCellValue("type");
wb.write(fileOut);
fileOut.close();
}
}
}catch (Exception e){
e.printStackTrace();
}finally {
}
}
//市级页面
public static void jsoupList2(String url,String provinceName,String provinceCode,SXSSFSheet sheet) throws Exception {
String cityName = "";
String cityCode = "";
url2 = url.replace(".html","");
Document document = Jsoup.connect(url).get();
Elements elements = document.select("body > table:nth-child(3) > tbody > tr:nth-child(1) > td > table > tbody > tr:nth-child(2) > td > table > tbody > tr > td > table > tbody");
Elements elements1 = elements.select("tbody > tr > td");
//j从2开始是因为他有个表头 统计用区划代码 名称
for (int j = 2; j < elements1.size(); j++) {
System.out.println("**********************i********************:"+i);
if(i%1000==0){
Thread.sleep(1000*60*10);
}
// Thread.sleep(500);
//判断是否是超链接，不是超链接也要获取数据
if(elements1.get(j).select("td > a").toString().equals("")){
String text = elements1.get(j).text();
if (j % 2 == 0) {
System.out.println("市代码:" + text);
sheet.createRow(i).createCell(0).setCellValue(text);
} else {
System.out.println("市名称:" + text);
sheet.getRow(i).createCell(1).setCellValue(text);
sheet.getRow(i).createCell(2).setCellValue(provinceCode);
sheet.getRow(i).createCell(3).setCellValue("3");
i++;
}
}else {
Elements elements2 = elements1.get(j).select("td > a");
for (int j1 = 0; j1 < elements2.size(); j1++) {
String text = elements2.get(j1).text();
if (j % 2 == 0) {
System.out.println("市代码:" + text);
cityCode = text;
sheet.createRow(i).createCell(0).setCellValue(text);
} else {
System.out.println("市名称:" + text);
cityName = text;
sheet.getRow(i).createCell(1).setCellValue(text);
sheet.getRow(i).createCell(2).setCellValue(provinceCode);
sheet.getRow(i).createCell(3).setCellValue("2");
i++;
String s = elements2.get(j1).select("a").attr("href");
//TODO 这里排除了海南的几个市区,更改为不排除任何市区
if(true) {
try {
jsoupList3(url1 + "/" + s,cityName,cityCode,provinceName,provinceCode, sheet);
} catch (SocketTimeoutException e) {
e.printStackTrace();
jsoupList3(url1 + "/" + s,cityName,cityCode,provinceName,provinceCode, sheet);
} catch (ConnectException e) {
e.printStackTrace();
jsoupList3(url1 + "/" + s,cityName,cityCode,provinceName,provinceCode, sheet);
}
}
}
}
}
}
}
//县级页面
public static void jsoupList3(String url,String cityName,String cityCode,String provinceName,String provinceCode,SXSSFSheet sheet) throws Exception {
Document document = Jsoup.connect(url).get();
Elements elements = document.select("body > table:nth-child(3) > tbody > tr:nth-child(1) > td > table > tbody > tr:nth-child(2) > td > table > tbody > tr > td > table > tbody");
Elements elements1 = elements.select("tbody > tr > td");
String xianName = "";
String xianCode = "";
//j从2开始是因为他有个表头 统计用区划代码 名称
for (int j = 2; j < elements1.size(); j++) {
System.out.println("**********************i********************:"+i);
if(i%1000==0){
Thread.sleep(1000*60*10);
}
// Thread.sleep(500);
//判断是否是超链接，不是超链接也要获取数据
if(elements1.get(j).select("td > a").toString().equals("")){
String text = elements1.get(j).text();
if (j % 2 == 0) {
System.out.println("县代码:" + text);
sheet.createRow(i).createCell(0).setCellValue(text);
} else {
System.out.println("县名称:" + text);
sheet.getRow(i).createCell(1).setCellValue(text);
sheet.getRow(i).createCell(2).setCellValue(cityCode);
sheet.getRow(i).createCell(3).setCellValue("3");
i++;
}
}else {
Elements elements2 = elements1.get(j).select("td > a");
for (int j1 = 0; j1 < elements2.size(); j1++) {
String text = elements2.get(j1).text();
xianName = text;
if (j % 2 == 0) {
xianCode = text;
System.out.println("县代码:" + xianCode);
sheet.createRow(i).createCell(0).setCellValue(text);
} else {
System.out.println("县名称:" + text);
sheet.getRow(i).createCell(1).setCellValue(text);
sheet.getRow(i).createCell(2).setCellValue(cityCode);
sheet.getRow(i).createCell(3).setCellValue("3");
i++;
String s = elements2.get(j1).select("a").attr("href");
try {
jsoupList4(url1 + "/" +provinceCode.substring(0,2)+"/"+ s,xianName,xianCode,cityName,cityCode,provinceName,provinceCode, sheet);
} catch (SocketTimeoutException e) {
e.printStackTrace();
jsoupList4(url1 + "/" +provinceCode.substring(0,2)+"/"+ s,xianName,xianCode,cityName,cityCode,provinceName,provinceCode, sheet);
} catch (ConnectException e) {
e.printStackTrace();
jsoupList4(url1 + "/" +provinceCode.substring(0,2)+"/"+ s,xianName,xianCode,cityName,cityCode,provinceName,provinceCode, sheet);
}
}
}
}
}
}
//街道页面
public static void jsoupList4(String url,String xianName,String xianCode,String cityName,String cityCode,String provinceName,String provinceCode,SXSSFSheet sheet) throws Exception {
Document document = Jsoup.connect(url).get();
Elements elements = document.select("body > table:nth-child(3) > tbody > tr:nth-child(1) > td > table > tbody > tr:nth-child(2) > td > table > tbody > tr > td > table > tbody");
Elements elements1 = elements.select("tbody > tr > td");
//j从2开始是因为他有个表头 统计用区划代码 名称
for (int j = 2; j < elements1.size(); j++) {
System.out.println("**********************i********************:"+i);
if(i%1000==0){
Thread.sleep(1000*60*10);
}
// Thread.sleep(500);
//判断是否是超链接，不是超链接也要获取数据
if(elements1.get(j).select("td > a").toString().equals("")){
String text = elements1.get(j).text();
if (j % 2 == 0) {
System.out.println("街道代码:" + text);
sheet.createRow(i).createCell(0).setCellValue(text);
} else {
System.out.println("街道名称:" + text);
sheet.getRow(i).createCell(1).setCellValue(text);
sheet.getRow(i).createCell(2).setCellValue(xianCode);
sheet.getRow(i).createCell(3).setCellValue("4");
i++;
}
}else {
Elements elements2 = elements1.get(j).select("td > a");
for (int j1 = 0; j1 < elements2.size(); j1++) {
String text = elements2.get(j1).text();
if (j % 2 == 0) {
System.out.println("街道代码:" + text);// TODO 这里不能截取，不然街道界别数据截不全
sheet.createRow(i).createCell(0).setCellValue(text);
} else {
System.out.println("街道名称:" + text);
sheet.getRow(i).createCell(1).setCellValue(text);
sheet.getRow(i).createCell(2).setCellValue(xianCode);
sheet.getRow(i).createCell(3).setCellValue("4");
i++;
}
}
}
}
}
}