高德地图搜索poi的api介绍地址
当前想法是爬取目标区域（作者所在小县城）的所有poi数据，存到数据库中作为原始数据，然后供其它系统调用，因为之前爬取过百度地图的poi数据，所以这次工作就驾轻就熟了。
1、首先注册一个高德地图的开发者账号，申请一个绑定Web服务的key，然后把刚注册的开发者账号认证一下： 申请账号、key就不赘述了，去高德地图开发平台很简单就能完成了，将账号认证是为了提高每日访问高德地图api接口的次数限制和并发请求。
2、根据上方api地址里面的介绍，总共分为4中搜索： 关键字搜索：通过用POI的关键字进行条件搜索，例如：肯德基、朝阳公园等；同时支持设置POI类型搜索，例如：银行 周边搜索：在用户传入经纬度坐标点附近，在设定的范围内，按照关键字或POI类型搜索； 多边形搜索：在多边形区域内进行搜索 ID查询：通过POI ID，查询某个POI详情，建议可同输入提示API配合使用
我的目标是某个区域的所有poi，所以选择的第三种：多边形搜索
3、多边形搜索最重要的参数就是polygon-》经纬度坐标对，我在百度地图坐标拾取系统拾取了我的目标区域的经纬度坐标对，如下图： ![百度地图坐标拾取系统]()
3步准备工作到这里就差不多结束了，在正式开始码代码之前先做个测试吧，用浏览器直接访问接口看看返回的数据（当然，高德的api接口有返回数据说明）
![api返回数据1]()
如上图，这里比较重要的一个属性是count，根据api的介绍count是搜索方案数目(最大值为1000)，所以说每次请求都会返回当前所搜所包含的poi个数，而大于1000的poi是没有办法获取到的。那么我如果想查询某个区域的全部数据，可以将这个区域再划分成更小的区域（显然是个递归操作）的集合，然后把这几个可以查到所有poi的区域的所有poi数据结合起来就是我最终需要的数据。可能口述不明朗，可以见下方草图：
![划分区域草图]()
好，可以开始撸代码了：
因为，整个调用API的过程都离不开经纬度，所以首先定义一个经纬度描述的类 `
//矩形块的经纬度标识， 左上角的经纬度 和右下角的经纬度
class RectangleCoordinate {
/**
* 矩形左上角经度
*/
private double x0;
/**
* 矩形左上角纬度
*/
private double y0;
/**
* 矩形右下角经度
*/
private double x1;
/**
* 矩形右下角纬度
*/
private double y1;
public RectangleCoordinate(double x0, double y0, double x1, double y1) {
this.x0 = x0;
this.y0 = y0;
this.x1 = x1;
this.y1 = y1;
}
/**
* [@return](https://my.oschina.net/u/556800) 获取矩形中心线的纬度
*/
public double getAverageY() {
return (y0 + y1) / 2;
}
/**
* [@return](https://my.oschina.net/u/556800) 获取矩形中心线的经度
*/
public double getAverageX() {
return (x0 + x1) / 2;
}
public double getX0() {
return x0;
}
public void setX0(double x0) {
this.x0 = x0;
}
public double getY0() {
return y0;
}
public void setY0(double y0) {
this.y0 = y0;
}
public double getX1() {
return x1;
}
public void setX1(double x1) {
this.x1 = x1;
}
public double getY1() {
return y1;
}
public void setY1(double y1) {
this.y1 = y1;
}
[@Override](https://my.oschina.net/u/1162528)
public String toString() {
return x0 + "," + y0 + "|" + x1 + "," + y1;
}
}`
然后需要一个调用api，获取返回数据的方法，这个方法参数就是矩形块，当然还需要一个页数，即当前方法获取的是某个矩形区域的第X页的数据（每页上线25个poi，默认20个poi）
/**
* @return 获取矩形块的poi数据
*/
private JSONObject getSearchResult(RectangleCoordinate coordinate, int page) {
RestTemplate restTemplate = new RestTemplate();
String url = getRequestGaodeUrl(coordinate,page);
String result = restTemplate.getForObject(url, String.class);
try {
try {
Thread.sleep(50);
} catch (InterruptedException e) {
e.printStackTrace();
}
return JSONObject.parseObject(result);
} catch (Exception e) {
logger.error("an error occurred when getting response of gaode map data for coordinate:[{}]", coordinate.toString());
}
return null;
}
当然，上方已经说过，如果矩形块返回数据count=1000，就说明当前矩形块需要分割，我的想法比较简单，将矩形块按照上方草图，在水平中心和垂直分心分割，1个矩形块就分割成4个小矩形块了，方法如下：
/**
* @return 将矩形4等分成小矩形 然后返回4个 小矩形的经纬度集合
*/
private List<RectangleCoordinate> getSplitRectangleList(RectangleCoordinate coordinate) {
List<RectangleCoordinate> splitRectangleList = new LinkedList<>();
splitRectangleList.add(new RectangleCoordinate(coordinate.getX0(), coordinate.getY0(), coordinate.getAverageX(), coordinate.getAverageY()));
splitRectangleList.add(new RectangleCoordinate(coordinate.getAverageX(), coordinate.getY0(), coordinate.getX1(), coordinate.getAverageY()));
splitRectangleList.add(new RectangleCoordinate(coordinate.getX0(), coordinate.getAverageY(), coordinate.getAverageX(), coordinate.getY1()));
splitRectangleList.add(new RectangleCoordinate(coordinate.getAverageX(), coordinate.getAverageY(), coordinate.getX1(), coordinate.getY1()));
return splitRectangleList;
}
目前，可以获取到矩形区域经纬度对的集合了，也有获取api数据的方法了，然后就是遍历页数获取数据，自定义操作数据。 当某次分页请求返回的poi个数小于每页最大个数的时候就认为当前区域poi已经完全请求到了。
private void startAnaMainGaode(RectangleCoordinate coordinate) throws AnalysisException {
//当前爬取的数据的页数索引
int page_num = 0;
//当前爬取内容是否是最后一页
boolean isLastPage = false;
JSONObject searchResult;
JSONArray datas = null;
logger.info("ready to analysis coordinate:[{}]", coordinate.toString());
while (!isLastPage) {
logger.info("is going to get data for page_" + page_num);
try {
searchResult = getSearchResult(coordinate, page_num);
datas = searchResult.getJSONArray("pois");
} catch (Exception e) {
logger.error("an error occurred when getting response of gaode map data for coordinate:[{}]", coordinate.toString());
}
if (datas != null && datas.size() < 20) {
isLastPage = true;
logger.info("get result counts is [{}], now page index is [{}]", datas.size(), page_num);
}
saveIntoDbGaode(datas);
page_num++;
}
}
private void saveIntoDbGaode(JSONArray result) {
JSONObject resultItem;
for (int i = 0; i < result.size(); i++) {
resultItem = result.getJSONObject(i);
try {
results.add(getInsertUnitObject(resultItem));
} catch (Exception e) {
logger.error("生成数据时异常，e: {}", e.getMessage());
e.printStackTrace();
}
}
if (results.size() > BATCHINSERTLIMIT || ISLAST) {
logger.info("is ready to batch insert into unit, total count is {}", results.size());
try {
dao.batchAddUnitGaode(results);
} catch (Exception e) {
logger.error("更新数据库异常，e: {}", e.getMessage());
}
results = new JSONArray();
}
}`
到此，基本方法都介绍过了，全部代码如下（因为都是简单方法和逻辑，不明白的留言交流）
//请求入口
public void GaodePoiSearch() {
//徐水区 final RectangleCoordinate searchAreaCoordinate = new RectangleCoordinate(115.521773, 39.106335, 115.801182, 38.943988);
//保定市
//final RectangleCoordinate searchAreaCoordinate = new RectangleCoordinate(114.332719,39.574064, 116.588688,38.179144);
List<RectangleCoordinate> validCoordinate = getValidCoordinate(searchAreaCoordinate);
logger.info("get all valid coordinate,size is [{}]", validCoordinate.size());
/**
* 获取到所有的小方块之后可以做一些处理， 比如存储到某个地方，以防发生异常，方便后面重新遍历，我这里暂未做处理
*/
validCoordinate.forEach(coor -> {
try {
startAnaMainGaode(coor);
} catch (AnalysisException e) {
e.printStackTrace();
}
});
ISLAST = true;
saveIntoDbGaode(new JSONArray());
}
/**
* [@return](https://my.oschina.net/u/556800) 获取矩形块中 符合 调用api的 小矩形块的集合
* 因为高德地图某个矩形块只能获取前1000条，所以要将矩形块分割成可以获取到全部数据的矩形块
* 如果当前矩形块请求数据返回的count<1000 即为符合条件的，否则将矩形块4等分 然后递归
*/
private List<RectangleCoordinate> getValidCoordinate(RectangleCoordinate coordinate) {
List<RectangleCoordinate> validCoordinate = new LinkedList<>();
JSONObject searchResult = getSearchResult(coordinate, 0);
if (searchResult.getIntValue("count") >= 1000) {
List<RectangleCoordinate> splitRectangleList = getSplitRectangleList(coordinate);
splitRectangleList.forEach(coor -> validCoordinate.addAll(getValidCoordinate(coor)));
} else {
logger.info("add a valid coordinate [{}]", coordinate.toString());
validCoordinate.add(coordinate);
}
return validCoordinate;
}
/**
* [@return](https://my.oschina.net/u/556800) 将矩形4等分成小矩形 然后返回4个 小矩形的经纬度集合
*/
private List<RectangleCoordinate> getSplitRectangleList(RectangleCoordinate coordinate) {
List<RectangleCoordinate> splitRectangleList = new LinkedList<>();
splitRectangleList.add(new RectangleCoordinate(coordinate.getX0(), coordinate.getY0(), coordinate.getAverageX(), coordinate.getAverageY()));
splitRectangleList.add(new RectangleCoordinate(coordinate.getAverageX(), coordinate.getY0(), coordinate.getX1(), coordinate.getAverageY()));
splitRectangleList.add(new RectangleCoordinate(coordinate.getX0(), coordinate.getAverageY(), coordinate.getAverageX(), coordinate.getY1()));
splitRectangleList.add(new RectangleCoordinate(coordinate.getAverageX(), coordinate.getAverageY(), coordinate.getX1(), coordinate.getY1()));
return splitRectangleList;
}
/**
* @return 获取矩形块的poi数据
*/
private JSONObject getSearchResult(RectangleCoordinate coordinate, int page) {
RestTemplate restTemplate = new RestTemplate();
String url = getRequestGaodeUrl(coordinate,page);
String result = restTemplate.getForObject(url, String.class);
try {
try {
Thread.sleep(50);
} catch (InterruptedException e) {
e.printStackTrace();
}
return JSONObject.parseObject(result);
} catch (Exception e) {
logger.error("an error occurred when getting response of gaode map data for coordinate:[{}]", coordinate.toString());
}
return null;
}
private void startAnaMainGaode(RectangleCoordinate coordinate) throws AnalysisException {
//当前爬取的数据的页数索引
int page_num = 0;
//当前爬取内容是否是最后一页
boolean isLastPage = false;
JSONObject searchResult;
JSONArray datas = null;
logger.info("ready to analysis coordinate:[{}]", coordinate.toString());
while (!isLastPage) {
logger.info("is going to get data for page_" + page_num);
try {
searchResult = getSearchResult(coordinate, page_num);
datas = searchResult.getJSONArray("pois");
} catch (Exception e) {
logger.error("an error occurred when getting response of gaode map data for coordinate:[{}]", coordinate.toString());
}
if (datas != null && datas.size() < 20) {
isLastPage = true;
logger.info("get result counts is [{}], now page index is [{}]", datas.size(), page_num);
}
saveIntoDbGaode(datas);
page_num++;
}
}
private void saveIntoDbGaode(JSONArray result) {
JSONObject resultItem;
for (int i = 0; i < result.size(); i++) {
resultItem = result.getJSONObject(i);
try {
results.add(getInsertUnitObject(resultItem));
} catch (Exception e) {
logger.error("生成数据时异常，e: {}", e.getMessage());
e.printStackTrace();
}
}
if (results.size() > BATCHINSERTLIMIT || ISLAST) {
logger.info("is ready to batch insert into unit, total count is {}", results.size());
try {
dao.batchAddUnitGaode(results);
} catch (Exception e) {
logger.error("更新数据库异常，e: {}", e.getMessage());
}
results = new JSONArray();
}
}
private JSONObject getInsertUnitObject(JSONObject resultItem) {
JSONObject unitDataObject = new JSONObject();
unitDataObject.put("uid", resultItem.getString("id"));
unitDataObject.put("name", resultItem.getString("name"));
unitDataObject.put("type", resultItem.getString("type"));
unitDataObject.put("tag", resultItem.getString("type"));
unitDataObject.put("address", resultItem.getString("address"));
unitDataObject.put("province", resultItem.getString("pname"));
unitDataObject.put("city", resultItem.getString("cityname"));
unitDataObject.put("area", resultItem.getString("adname"));
String tel = resultItem.getString("tel");
if (tel != null && !"[]".equals(tel)) {
unitDataObject.put("telephone", tel);
}
try {
JSONArray url = resultItem.getJSONArray("website");
if (url != null && url.size() > 0) {
unitDataObject.put("detail_url", url.getString(0));
}
} catch (Exception e) {
unitDataObject.put("detail_url", resultItem.getString("website"));
}
JSONArray photos = resultItem.getJSONArray("photos");
if (photos != null && photos.size() > 0) {
StringBuilder images = new StringBuilder();
for (int j = 0; j < photos.size(); j++) {
images.append(j == 0 ? "" : ";").append(photos.getJSONObject(j).getString("url"));
}
unitDataObject.put("images", images.toString());
}
String entr_location = resultItem.getString("location");
if (StringUtils.isEmpty(entr_location)) {
entr_location = resultItem.getString("entr_location");
}
if (!StringUtils.isEmpty(entr_location)) {
unitDataObject.put("lng", entr_location.split(",")[0]);
unitDataObject.put("lat", entr_location.split(",")[1]);
}
return unitDataObject;
}
private String getRequestGaodeUrl(RectangleCoordinate coordinate, int page) {
return "https://restapi.amap.com/v3/place/polygon?" +
"key=xxxxxxxxxxxxxxxxxxxxxxx&polygon=" + coordinate.toString() + "&page=" + page + "&types=010000|" +
"010100|010101|010102|010103|010104|010105|010107|010108|010109|010110|010111|010112|010200|010300|010400|" +
"010401|010500|010600|010700|010800|010900|010901|011000|011100|020000|020100|020101|020102|020103|020104|" +
"020105|020106|020200|020201|020202|020203|020300|020301|020400|020401|020402|020403|020404|020405|020406|" +
"020407|020408|020600|020601|020602|020700|020701|020702|020703|020800|020900|020904|020905|021000|021001|" +
"021002|021003|021004|021100|021200|021201|021202|021203|021300|021301|021400|021401|021500|021501|021600|" +
"021601|021602|021700|021701|021702|021800|021802|021803|021804|021900|022000|022100|022200|022300|022301|" +
"022400|022500|022501|022502|022600|022700|022800|022900|023000|023100|023200|023300|023301|023400|023500|" +
"025000|025100|025200|025300|025400|025500|025600|025700|025800|025900|026000|026100|026200|026300|029900|" +
"030000|030100|030200|030201|030202|030203|030204|030205|030206|030300|030301|030302|030303|030400|030401|" +
"030500|030501|030502|030503|030504|030505|030506|030507|030508|030700|030701|030702|030800|030801|030802|" +
"030803|030900|031000|031004|031005|031100|031101|031102|031103|031104|031200|031300|031301|031302|031303|" +
"031400|031401|031500|031501|031600|031601|031700|031701|031702|031800|031801|031802|031900|031902|031903|" +
"031904|032000|032100|032200|032300|032400|032401|032500|032600|032601|032602|032700|032800|032900|033000|" +
"033100|033200|033300|033400|033401|033500|033600|035000|035100|035200|035300|035400|035500|035600|035700|" +
"035800|035900|036000|036100|036200|036300|039900|040000|040100|040101|040200|040201|050000|050100|050101|" +
"050102|050103|050104|050105|050106|050107|050108|050109|050110|050111|050112|050113|050114|050115|050116|" +
"050117|050118|050119|050120|050121|050122|050123|050200|050201|050202|050203|050204|050205|050206|050207|" +
"050208|050209|050210|050211|050212|050213|050214|050215|050216|050217|050300|050301|050302|050303|050304|" +
"050305|050306|050307|050308|050309|050310|050311|050400|050500|050501|050502|050503|050504|050600|050700|" +
"050800|050900|060000|060100|060101|060102|060103|060200|060201|060202|060300|060301|060302|060303|060304|" +
"060305|060306|060307|060308|060400|060401|060402|060403|060404|060405|060406|060407|060408|060409|060411|" +
"060413|060414|060415|060500|060501|060502|060600|060601|060602|060603|060604|060605|060606|060700|060701|" +
"060702|060703|060704|060705|060706|060800|060900|060901|060902|060903|060904|060905|060906|060907|061000|" +
"061001|061100|061101|061102|061103|061104|061200|061201|061202|061203|061204|061205|061206|061207|061208|" +
"061209|061210|061211|061212|061213|061214|061300|061301|061302|061400|061401|070000|070100|070200|070201|" +
"070202|070203|070300|070301|070302|070303|070304|070305|070306|070400|070401|070500|070501|070600|070601|" +
"070603|070604|070605|070606|070607|070608|070609|070610|070700|070701|070702|070703|070704|070705|070706|" +
"070800|070900|071000|071100|071200|071300|071400|071500|071600|071700|071800|071801|071900|071901|071902|" +
"071903|072000|072001|080000|080100|080101|080102|080103|080104|080105|080106|080107|080108|080109|080110|" +
"080111|080112|080113|080114|080115|080116|080117|080118|080119|080200|080201|080202|080300|080301|080302|" +
"080303|080304|080305|080306|080307|080308|080400|080401|080402|080500|080501|080502|080503|080504|080505|" +
"080600|080601|080602|080603|090000|090100|090101|090102|090200|090201|090202|090203|090204|090205|090206|" +
"090207|090208|090209|090210|090211|090300|090400|090500|090600|090601|090602|090700|090701|090702|100000|" +
"100100|100101|100102|100103|100104|100105|100200|100201|110000|110100|110101|110102|110103|110104|110105|" +
"110106|110200|110201|110202|110203|110204|110205|110206|110207|110208|110209|120000|120100|120200|120201|" +
"120202|120203|120300|120301|120302|120303|120304|130000|130100|130101|130102|130103|130104|130105|130106|" +
"130107|130200|130201|130202|130300|130400|130401|130402|130403|130404|130405|130406|130407|130408|130409|" +
"130500|130501|130502|130503|130504|130505|130506|130600|130601|130602|130603|130604|130605|130606|130700|" +
"130701|130702|130703|140000|140100|140101|140102|140200|140201|140300|140400|140500|140600|140700|140800|" +
"140900|141000|141100|141101|141102|141103|141104|141105|141200|141201|141202|141203|141204|141205|141206|" +
"141207|141300|141400|141500|150000|150100|150101|150102|150104|150105|150106|150107|150200|150201|150202|" +
"150203|150204|150205|150206|150207|150208|150209|150210|150300|150301|150302|150303|150304|150400|150500|" +
"150501|150600|150700|150701|150702|150703|150800|150900|150903|150904|150905|150906|150907|150908|150909|" +
"151000|151100|151200|151300|160000|160100|160101|160102|160103|160104|160105|160106|160107|160108|160109|" +
"160110|160111|160112|160113|160114|160115|160117|160118|160119|160120|160121|160122|160123|160124|160125|" +
"160126|160127|160128|160129|160130|160131|160132|160133|160134|160135|160136|160137|160138|160139|160140|" +
"160141|160142|160143|160144|160145|160146|160147|160148|160149|160150|160151|160152|160200|160300|160301|" +
"160302|160303|160304|160305|160306|160307|160308|160309|160310|160311|160312|160314|160315|160316|160317|" +
"160318|160319|160320|160321|160322|160323|160324|160325|160326|160327|160328|160329|160330|160331|160332|" +
"160333|160334|160335|160336|160337|160338|160339|160340|160341|160342|160343|160344|160345|160346|160347|" +
"160348|160349|160400|160401|160402|160403|160404|160405|160406|160407|160408|160500|160501|160600|170000|" +
"170100|170200|170201|170202|170203|170204|170205|170206|170207|170208|170209|170300|170400|170401|170402|" +
"170403|170404|170405|170406|170407|170408|180000|180100|180101|180102|180103|180104|180200|180201|180202|" +
"180203|180300|180301|180302|180400|180500|190000|190100|190101|190102|190103|190104|190105|190106|190107|" +
"190108|190109|190200|190201|190202|190203|190204|190205|190300|190301|190302|190303|190304|190305|190306|" +
"190307|190308|190309|190310|190311|190400|190401|190402|190403|190500|190600|190700|200000|200100|200200|" +
"200300|200301|200302|200303|200304|200400|220000|220100|220101|220102|220103|220104|220105|220106|220107|" +
"220200|220201|220202|220203|220204|220205|970000|990000|991000|991001|991400|991401|991500&extensions=all";
}
/**
* 矩形块的经纬度标识， 左上角的经纬度 和右下角的经纬度
*/
class RectangleCoordinate {
/**
* 矩形左上角经度
*/
private double x0;
/**
* 矩形左上角纬度
*/
private double y0;
/**
* 矩形右下角经度
*/
private double x1;
/**
* 矩形右下角纬度
*/
private double y1;
public RectangleCoordinate(double x0, double y0, double x1, double y1) {
this.x0 = x0;
this.y0 = y0;
this.x1 = x1;
this.y1 = y1;
}
/**
* @return 获取矩形中心线的纬度
*/
public double getAverageY() {
return (y0 + y1) / 2;
}
/**
* @return 获取矩形中心线的经度
*/
public double getAverageX() {
return (x0 + x1) / 2;
}
public double getX0() {
return x0;
}
public void setX0(double x0) {
this.x0 = x0;
}
public double getY0() {
return y0;
}
public void setY0(double y0) {
this.y0 = y0;
}
public double getX1() {
return x1;
}
public void setX1(double x1) {
this.x1 = x1;
}
public double getY1() {
return y1;
}
public void setY1(double y1) {
this.y1 = y1;
}
@Override
public String toString() {
return x0 + "," + y0 + "|" + x1 + "," + y1;
}
}`
更新（2018-09-20）：
1、时间问题，当前50ms请求一次api接口，跑完小县城的数据（几万条）大概需要十分钟左右吧，把整个市区主要数据跑完断断续续的用了一天吧，最后跑了近27W数据
![]()
2、应用问题，原本的想法就是做个简单的小程序，把跑来的数据加以利用，做个电话本类似的应用，具体可以扫下方小程序码体验
![]()
更新（2019-01-28）：
有一些朋友向我要源码，可能大多是新手，尽管思路给了，代码还是写不出来。其实上方我把主要的代码基本都发布出来了，但是应各位要求，我把源码提交到github了，可以访问 我的github 查看
更新（2019-07-24）
想到一个弊端，并找到了解决方法：
很多朋友使用上文提供的方法时，难免会得到一些”垃圾数据“，何为垃圾数据呢？比如我爬取保定的某些数据，开始大致选了一个区域，为了爬取到所有的数据，就要保证所选区域要涵盖保定，最后爬到的数据就不止保定的数据了，其他区域的数据就为垃圾数据，如下图：
![]()
看到没有，在尽可能小的区域内，垃圾数据所在区域也几乎占了小一半了，除了临近的市区（任丘等），也包含了其他省（山西，北京等）的数据。除了区域不精准，更可怕的是像北京这种大城市，poi数量很大，所以会造成爬取的数据可能只有部分是目标数据。
有人会说，可以在爬完数据之后再处理。也不是不可以，不过这个过程中调用api，io操作费时费力。
当前我想到的办法是，根据原文方法拿到要爬取的区域之后，先判断是否是我们想要的区域（方形区域4个点至少有一个在目标区域内），否则就舍弃掉。比如上图右上角的区域，拿到之后发现都是北京的，舍弃。
具体方法还要调用百度或高德提供的逆地理编码接口。点这里看介绍。
根据区域的location，调用接口，得到返回的数据中会包含该location的country、province、city等。然后进行过滤就OK了。
优化核心类似：之前目标区域划分成的小矩形块有10万个，根据逆地理编码接口，过滤掉其他省市，剩余5千个，获取这5千个小矩形块的poi数据即可。
我当前的应用中，因为涉及到北京市，所以划分的小矩形块和目标小矩形块差异较大，类似下图： ![]()
