最近的工作是带一个 商城项目，客户需求 需要圈选访问商城的会员，比如访问了A页面，没有访问B页面，或者是同时访问A页面，并且触发了A1事件的 等等信息或者是url加有标识，从某一外部资源访问 进来的会员是哪些。
基于市面的统计软件，比如 百度统计，站长统计，腾讯统计 等等都无法满足业务，不能圈选。所以自建了一套系统，来做会员圈选
前端收集数据js
<script>
var _wx_tongji_={};
_wx_tongji_.addEvent=function(ev,type,val){
var formData = new FormData();
formData.append('ev',ev);
formData.append('type',type);
formData.append('val',val||0);
formData.append('url',encodeURI(location.href));
//创建xhr对象
var xhr = new XMLHttpRequest();
//设置xhr请求的超时时间
xhr.timeout = 3000;
//设置响应返回的数据格式
xhr.responseType = "text";
//创建一个 post 请求，采用异步
xhr.open('POST', ctx+'/statistics/tongji.do', true);
//設置 header X-Requested-With:XMLHttpRequest
xhr.setRequestHeader("X-Requested-With", "XMLHttpRequest");
//注册相关事件回调处理函数
xhr.onload = function(e) {
if(this.status == 200||this.status == 304){
console.log(e.target.responseText);
}
};
//发送数据
xhr.send(formData);
};
function getUrlParam(name) {
var reg = new RegExp("(^|&)" + name + "=([^&]*)(&|$)", "i");
var r = window.location.search.substr(1).match(reg);
if (r != null) return decodeURI(r[2]); return null;
}
(function(){
var wxrw=getUrlParam("from");
if(wxrw&&wxrw.startsWith("wxrw")){
_wx_tongji_.addEvent("access","wxrw",wxrw);
}
})();
</script>前端页面 通过 _wx_tongji_,addEvent() 来提交访问以及事件数据
后台接受数据，以logback形式 写入日志
public String tongji(HttpServletRequest request){
String url=EncodeUtils.urlDecode(EncodeUtils.urlDecode(request.getParameter("url"))).replaceAll(" ", "");
String ev=request.getParameter("ev");
String type=request.getParameter("type");
String val=request.getParameter("val");
String mobile="0";
String openid=ServletUtils.getOpenId();
if(StringUtils.isBlank(openid))
openid="0";
JSONObject user=ServletUtils.getUser();
if(user!=null){
mobile=user.getString("mobile");
}
StringBuffer str=new StringBuffer();
str.append(ev);
str.append(" ");
str.append(type);
str.append(" ");
str.append(val);
str.append(" ");
str.append(openid);
str.append(" ");
str.append(mobile);
str.append(" ");
str.append(url);
logger.info(str.toString());
return null;
}日志记录数据格式如下
年月日 时分秒 访问/事件 类型 值 ID 手机号 具体访问url
170719 18:22:19 access goods 8498 id1111 手机号11111 https://url11111
170719 18:22:19 access goods 8081 id2222 手机号22222 https://url22222
上图为 前台页面查询 输入 以及反馈结果，数字为 对应子对象的数据处理结果。
前台页面提交后台的参数为json,提交的时候 result 以及result_value 值为空， 这2个值为后台执行分析之后的反馈结果记录
{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.6688393061188297.log",
"flag":0,
"result_value":"278674",
"sons":[
{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9578616973272448.log",
"search":{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.34378951503811495.log",
"flag":0,
"start":"201706091948",
"result_value":"287696",
"sons":[
{
"val":"9235",
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.2874368262364775.log",
"ev":"access",
"result_value":"9061",
"type":"goods"
},
{
"val":"1188",
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.10730146303857035.log",
"ev":"access",
"result_value":"286835",
"type":"page"
}
],
"end":"201707191948"
},
"result_value":"278635",
"filter":{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9342168703806003.log",
"flag":0,
"result_value":"9061",
"sons":[
{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9342168703806003.log",
"flag":1,
"start":"201707042020",
"result_value":"9061",
"sons":[
{
"val":"9235",
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9342168703806003.log",
"ev":"access",
"result_value":"9061",
"type":"goods"
}
],
"end":"201707192020"
}
]
}
},
{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.027657177982776227.log",
"search":{
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.027657177982776227.log",
"flag":1,
"result_value":"223",
"start":"201707171951",
"sons":[
{
"val":"9699",
"result":"/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.027657177982776227.log",
"ev":"access",
"result_value":"223",
"type":"goods"
}
],
"end":"201707191951"
},
"result_value":"223",
"filter":{
"flag":0,
"result_value":"",
"sons":[
]
}
}
],
"html":"https://abc//tp20170719203249/201707192032490.6688393061188297"
}后台接受参数解析处理代码
public static String search(String searchValueStr){
return service.toSearch(searchValueStr);
}
String toSearch(String searchValueStr){
//查询条件
JSONObject searchValues=JSONObject.parseObject(searchValueStr);
//记录所有需要执行的命令集合
List<String> command=new ArrayList<>();
//临时目录
String _tempPath=searchPath+"/tp"+new SimpleDateFormat("yyyyMMddHHmmss").format(new Date());
String result=m_search(searchValues, command, _tempPath,null);
searchValues.put("result", result);
//复制一份数据 生成html文件
String cp_command=" gawk '{print $0,\" </br>\"}' "+result+" > "+result.replace(".log", ".html");
command.add(cp_command);
//统计所有分析结果的数据 行数
String wc_l_command=" wc -l "+_tempPath+"/*.log |gawk '{print $2,\"=\",$1}'> "+_tempPath+"/result.properties";
command.add(wc_l_command);
//执行脚本
execCommand(_tempPath, command);
//设定每一个命令执行的结果得到的数据条目数 到结果集内
setResult(searchValues, getResultCount(_tempPath));
//设置反馈前台可以访问的最终结果的html地址
result=domainPath+result.replace(searchPath, "").replace(".log", "");
searchValues.put("html", result);
//返回结果
return searchValues.toJSONString();
}
void setResult(JSONObject params,Properties prop){
List<String> keys=new ArrayList<>();
keys.addAll(params.keySet());
for(String key:keys){
Object v=params.get(key);
if("result".equals(key)){
params.put("result_value", prop.getProperty(v.toString(),"0"));
}else if(v instanceof JSONArray){
JSONArray value=(JSONArray) v;
for(int i=0;i<value.size();i++)
setResult(value.getJSONObject(i), prop);
}else if (v instanceof JSONObject){
setResult((JSONObject) v, prop);
}
}
}
Properties getResultCount(String path) {
Properties props = new Properties();
InputStream is = null;
try {
InputStream in = new BufferedInputStream(new FileInputStream(new File(path+"/result.properties")));
props.load(in);
} catch (Exception e) {
e.printStackTrace();
} finally {
IOUtils.closeQuietly(is);
}
for(Object key:props.keySet())
System.out.println(key.toString()+":"+props.getProperty(key.toString()));
System.out.println("props");
return props;
}
String m_search(JSONObject search,List<String> command,String path,JSONObject parent){
//获取查询条件集合
JSONArray sons=search.getJSONArray("sons");
JSONObject _search=search.getJSONObject("search");
//如果沒有子对象，并且没有搜索对象，则标识是最底层的 条件查找，并非上层的组合条件
if((sons==null||sons.size()==0)&&_search==null){
String _result=DataSearchUtil.sonSearch(search, command, path, parent);
search.put("result", _result);
return _result;
}
//如果没有子对象，则标识是子层的 检索，以及排除条件处理
if(sons==null||sons.size()==0){
String _result_search=m_search(_search, command, path,_search);
String _result_exclude="";
JSONObject _filter=search.getJSONObject("filter");
if(_filter==null){
_result_exclude= _result_search;
}else{
JSONArray _sons=_filter.getJSONArray("sons");
if(_sons==null||_sons.size()==0){
_result_exclude=_result_search;
}else{
String _result_filter=m_search(_filter, command, path,null);
_result_exclude=searchExclude(_result_search,_result_filter, command,path);
}
}
search.put("result", _result_exclude);
return _result_exclude;
}
// 如果存在子对象的，则标识需要 遍历查询子对象得到其 结果
//每一个子查询最终输出结果的文件名称集合
List<String> result_fileNames=new ArrayList<>();
//获取 查询模块的逻辑判断类型 且/或
boolean flag=1==search.getInteger("flag");
for(int j=0;j<sons.size();j++){
JSONObject s_search=sons.getJSONObject(j);
String result=m_search(s_search,command,path,search);
result_fileNames.add(result);
}
String res="";
if(result_fileNames.size()==1)
res= result_fileNames.get(0);
else
res=flag?searchAnd(result_fileNames, command,path):searchOr(result_fileNames, command,path);
search.put("result", res);
return res;
}
String searchOr(List<String> fileNames,List<String> command,String path){
String _command=" gawk '{a[$1]+=1}END{for(i in a) print i}' " ;
for(String result_file:fileNames)
_command+=result_file+" ";
String fileName=createFileName(path);
_command+= ">"+fileName;
command.add(_command);
return fileName;
}
String searchAnd(List<String> fileNames,List<String> command,String path){
// 如果是且的条件，则 加入判断 出现次数 等于 要查询的文件的数目，即为 所有文件内都存在的数据
String _command=" gawk '{a[$1]+=1}END{for(i in a) if(a[i]=="+fileNames.size()+") print i}' " ;
for(String result_file:fileNames)
_command+=result_file+" ";
String fileName=createFileName(path);
_command+= ">"+fileName;
command.add(_command);
return fileName;
}
String searchExclude(String search,String filter,List<String> command ,String path){
// gawk '{if(ARGIND==1){a[$0]}else{if($0 in a) delete a[$0]}}END{for (i in a) print i} a.log b.log c.log 从a里面 排除内容存在与 b c里面的条目
//判断 流对应文件是第一个，则放入集合，如果不是第一个并且 对应数据在集合内存在，则从集合内删除该条数据，最终打印输出结果
String command_filter=" gawk '{if(ARGIND==1){a[$0]}else{if($0 in a) delete a[$0]}}END{for (i in a) print i}' "+search+" "+ filter;
//查询 排除条件之后的结果输出文件
String fileName=createFileName(path);
command_filter+=" > "+fileName;
command.add(command_filter);
return fileName;
}
void execCommand(String path,List<String>command){
try {
//创建临时目录
new File(path).mkdir();
//创建命令文件
File file=new File(path+"/command.sh");
file.createNewFile();
//写入命令
FileWriter fw = new FileWriter(file.getAbsoluteFile());
BufferedWriter bw = new BufferedWriter(fw);
for(String com:command)
bw.write(com+"\n");
bw.close();
//赋予命令文件 可执行权限
file.setExecutable(true);
//执行命令文件
Process p = Runtime.getRuntime().exec(new String[]{"/bin/sh","-c",path+"/command.sh"});
InputStream is = p.getInputStream();
p.waitFor();
is.close();
p.destroy();
} catch (Exception e) {
e.printStackTrace();
}
}
public static String getCommandFind(String path,String start,String end){
// 根据 start end 更改 临时文件 start end 的 相关时间
// touch -t 201707070000 /var/log/tomcat/wx_tongji_temp/t_start;
// touch -t 201707100000 /var/log/tomcat/wx_tongji_temp/t_end;
// find . -name "*.log" -type f -newer /var/log/tomcat/wx_tongji_temp/t_start ! -newer /var/log/tomcat/wx_tongji_temp/t_end
String start_file=path+"/start_"+(int)(Math.random()*100000);
String end_file=path+"/end_"+(int)(Math.random()*100000);
String t_start=" touch -t "+start+" "+ start_file;
String t_end=" touch -t "+end+" "+ end_file;
return t_start+";\n"+t_end+";\n"+" find "+logPath+" -name '*.log' -type f -newer "+start_file+" ! -newer "+end_file;
}
public static void main(String[] args) {
TongjiService service=new TongjiService();
service.getResultCount("d:\\");
} public String sonSearch(JSONObject s_search, List<String> command, String path,JSONObject parent) {
// 获取开始时间
String start = parent.getString("start");
// 获取结束时间
String end = parent.getString("end");
// 根据时间范围 得到 find 命令
String command_find = TongjiService.getCommandFind(path, start, end);
String ev=s_search.getString("ev");
String type=s_search.getString("type");
String val=s_search.getString("val");
if("-1".equals(val))val="";
//如果是 且的关系，则分别独立查询
String fileName=TongjiService.createFileName(path);
// find . -name *.log |xargs grep ' access goods 1848' | gawk -F ' ' '{a[$6]}END{for(i in a) print i}' > a.log
String command_grep_gawk=command_find+" | xargs grep '"+ev+" "+type+(!"-1".equals(val)?" "+val:"")+" ' | gawk -F ' ' '{a[$6]}END{for(i in a) print i}' >"+fileName;
command.add(command_grep_gawk);
return fileName;
}整体逻辑，接受参数，解析内部逻辑得到所有需要 执行分析的命令，最终生成命令脚本文件，执行脚本。
某一个子查询的结果 如上面的 json 参数以及对应结果
总用量 42396
-rw-r--r--. 1 root root 6440 7月 19 20:32 201707192032490.027657177982776227.log
-rw-r--r--. 1 root root 8318105 7月 19 20:32 201707192032490.10730146303857035.log
-rw-r--r--. 1 root root 262742 7月 19 20:32 201707192032490.2874368262364775.log
-rw-r--r--. 1 root root 8343074 7月 19 20:32 201707192032490.34378951503811495.log
-rw-r--r--. 1 root root 10032154 7月 19 20:32 201707192032490.6688393061188297.html
-rw-r--r--. 1 root root 8081436 7月 19 20:32 201707192032490.6688393061188297.log
-rw-r--r--. 1 root root 262742 7月 19 20:32 201707192032490.9342168703806003.log
-rw-r--r--. 1 root root 8080332 7月 19 20:32 201707192032490.9578616973272448.log
-rwxr--r--. 1 root root 3465 7月 19 20:32 command.sh
-rw-r--r--. 1 root root 0 7月 19 19:51 end_50495
-rw-r--r--. 1 root root 0 7月 19 19:48 end_71400
-rw-r--r--. 1 root root 0 7月 19 19:48 end_96108
-rw-r--r--. 1 root root 0 7月 19 20:20 end_98721
-rw-r--r--. 1 root root 689 7月 19 20:32 result.properties
-rw-r--r--. 1 root root 0 7月 17 19:51 start_13790
-rw-r--r--. 1 root root 0 7月 4 20:20 start_21328
-rw-r--r--. 1 root root 0 6月 9 19:48 start_58068
-rw-r--r--. 1 root root 0 6月 9 19:48 start_76742 touch -t 201706091948 /var/log/tomcat/wx_tongji_search/tp20170719203249/start_58068;
touch -t 201707191948 /var/log/tomcat/wx_tongji_search/tp20170719203249/end_71400;
find /var/log/tomcat/wx_tongji -name '*.log' -type f -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/start_58068 ! -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/end_71400 | xargs grep 'access goods 9235 ' | gawk -F ' ' '{a[$6]}END{for(i in a) print i}' >/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.2874368262364775.log
touch -t 201706091948 /var/log/tomcat/wx_tongji_search/tp20170719203249/start_76742;
touch -t 201707191948 /var/log/tomcat/wx_tongji_search/tp20170719203249/end_96108;
find /var/log/tomcat/wx_tongji -name '*.log' -type f -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/start_76742 ! -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/end_96108 | xargs grep 'access page 1188 ' | gawk -F ' ' '{a[$6]}END{for(i in a) print i}' >/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.10730146303857035.log
gawk '{a[$1]+=1}END{for(i in a) print i}' /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.2874368262364775.log /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.10730146303857035.log >/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.34378951503811495.log
touch -t 201707042020 /var/log/tomcat/wx_tongji_search/tp20170719203249/start_21328;
touch -t 201707192020 /var/log/tomcat/wx_tongji_search/tp20170719203249/end_98721;
find /var/log/tomcat/wx_tongji -name '*.log' -type f -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/start_21328 ! -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/end_98721 | xargs grep 'access goods 9235 ' | gawk -F ' ' '{a[$6]}END{for(i in a) print i}' >/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9342168703806003.log
gawk '{if(ARGIND==1){a[$0]}else{if($0 in a) delete a[$0]}}END{for (i in a) print i}' /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.34378951503811495.log /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9342168703806003.log > /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9578616973272448.log
touch -t 201707171951 /var/log/tomcat/wx_tongji_search/tp20170719203249/start_13790;
touch -t 201707191951 /var/log/tomcat/wx_tongji_search/tp20170719203249/end_50495;
find /var/log/tomcat/wx_tongji -name '*.log' -type f -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/start_13790 ! -newer /var/log/tomcat/wx_tongji_search/tp20170719203249/end_50495 | xargs grep 'access goods 9699 ' | gawk -F ' ' '{a[$6]}END{for(i in a) print i}' >/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.027657177982776227.log
gawk '{a[$1]+=1}END{for(i in a) print i}' /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.9578616973272448.log /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.027657177982776227.log >/var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.6688393061188297.log
gawk '{print $0," </br>"}' /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.6688393061188297.log > /var/log/tomcat/wx_tongji_search/tp20170719203249/201707192032490.6688393061188297.html
wc -l /var/log/tomcat/wx_tongji_search/tp20170719203249/*.log |gawk '{print $2,"=",$1}'> /var/log/tomcat/wx_tongji_search/tp20170719203249/result.propertiestouch 根据时间生成 开始和结束文件
find -newer 对比开始 结束文件 来定位时间段内的日志文件
grep 来检索需要的内容
gawk 来格式化输出结果
gawk -F ' ' '{a[$6]}END{for(i in a) print i}'gawk '{if(ARGIND==1){a[$0]}else{if($0 in a) delete a[$0]}}END{for (i in a) print i}' a.log b.logARGIND==1 判断是第一个文件里面的内容 ，如果是 则放入数组 a内，如果不是，并且 在数组内存在，则从输入内删除， 达到目的 A文件 排除 存在于B 文件内的数据
gawk '{a[$1]+=1}END{for(i in a) if(a[i]=="+fileNames.size()+") print i}'
gawk '{a[$1]+=1}END{for(i in a) if(a[i]==2) print i}' a.log b.loga[$1]+=1 遍历数据 放入数组内，并且存在重复则数量+1
for(i in a) 遍历结果， if(a[i]==2) 如果数量==2 即== 要处理的文件的数量，则表示在2个文件内都存在，则输出内容，达到目的 获取2个文件重叠的内容
最后 将结果文件，复制一份生成html文件，加入</br> 作用于 html显示 换行
最后 wc -l 统计每个文件的条目数， gawk '{print $2,"=",$1}' 格式化输出变成 文件名称=数量 的properties 文件，用于解析。
实际存储文件，关联 nginx 目录，可以通过 http://a.com/***/ web来访问。
最终反馈的结果到前台页面，html 展示，log 用于下载。
扫一扫
2490
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
