需要解析一种类xml的数据文件，数据格式1如下：
<head> //文件头 <type>xtype</type> <condition> key1=value1 key2=value2 </condition> <mea> key3=value3 key4=value4 </mea> <xxxx>//多个 ... </xxx> </head> <data> //数据域，多个 phi rcs ang 1 2 3 2 3 4 </data> <data> phi rcs ang 3 4 5 4 5 6 </data>
数据格式2：
#xtype //comment [condition] 项1=值1 项2=值2 [/condition] [mea] key3=value3 key4=value4 [/mea] [data] phi rcs ang 1 1 1 2 2 2 [/data] [data] phi rcs ang 3 1 1 4 2 2 [/data]
该数据格式类似xml，我们需要解析的是head中的所有标签，及标签中的键值对（key=value），并将data域中的数据保存成浮点型数组。
采用类似xml的解析方式，递归进行解析
具体代码如下：
#ifndef MYPARSEGJBDATA_H_20170114 #define MYPARSEGJBDATA_H_20170114 #include <string> #include <map> #include <vector> #include <iostream> #include <fstream> #include <sstream> using namespace std; /*file format1: <head> <type>xtype</type> <conditon> key1=value1 key2=value2 </condition> <mea> key3=value3 key4=value4 </mea> </head> <data> phi rcs ang 1 1 1 2 2 2 </data> */ typedef struct stMyDataLabel{ string Label;//标签名称 vector<string> Content;//对于<lable>content</label>的形式，纯字符串，非键值对 vector<float> Values;//对应数据域，一维数组表示二维数组；数组的列数为Contens'size map<string,string> KeyVal;//<label>内的键值对 vector<stMyDataLabel> SubItems;//子节点列表 stMyDataLabel(){ Label = ""; Content.clear(); Values.clear(); KeyVal.clear(); SubItems.clear(); } void Clear(){//递归清空内存 for (vector<stMyDataLabel>::iterator itr = SubItems.begin();itr!=SubItems.end();itr++) { itr->Clear(); } Content.clear(); Values.clear(); KeyVal.clear(); SubItems.clear(); } friend ostream& operator <<(ostream& myout,stMyDataLabel& m_Data){//输出流 myout<<m_Data.Label<<endl; for (vector<string>::iterator sitr = m_Data.Content.begin(); sitr != m_Data.Content.end(); sitr++) { myout<<" "<<*sitr<<" "; } myout<<endl; for (auto mitr = m_Data.KeyVal.begin(); mitr != m_Data.KeyVal.end(); mitr++) { myout<<" "<<mitr->first<<" = "<<mitr->second<<" "; } myout<<endl; int dsize = m_Data.Content.size(); int curdidx = 0; for(vector<float>::iterator fitr = m_Data.Values.begin(); fitr != m_Data.Values.end(); fitr++){ myout<<" "<<*fitr<<" "; curdidx++; if(curdidx >= dsize){ cout<<endl; curdidx=0; } } myout<<endl; for (vector<stMyDataLabel>::iterator itr = m_Data.SubItems.begin();itr != m_Data.SubItems.end(); itr++) { myout<<*itr<<endl; } return myout; } }stMyDataLabel; class MyDataParse{//数据类型1的解析类 public: vector<stMyDataLabel> m_Data; stMyDataLabel m_Header; char m_StartLabel; char m_EndLabel; public: MyDataParse(){ m_StartLabel = '<'; m_EndLabel = '>'; } void Init(){ m_Header.Clear(); for (vector<stMyDataLabel>::iterator itr = m_Data.begin();itr!=m_Data.end();itr++) { itr->Clear(); } } void Print(ostream& myout){ myout<<m_Header<<endl; for (vector<stMyDataLabel>::iterator ditr = m_Data.begin();ditr!=m_Data.end();ditr++) myout<<*ditr<<endl; } bool Parse(char* filename){ if(filename == NULL) return false; ifstream myin(filename,ios::in); bool flag = ParseLabel(myin,m_Header); if(!flag) return false; while(!myin.eof()){ stMyDataLabel data; flag = ParseLabel(myin,data); m_Data.push_back(data); } return flag; } bool ParseLabel(istream& myin,stMyDataLabel& label){ string str; bool flag = true; do{ myin >> str; int sidx=0,eidx=0; sidx = str.find(m_StartLabel); eidx = str.find(m_EndLabel); if(sidx >= 0 && eidx >= 0 && eidx > sidx){//<lable> or </label> if(str.at(sidx+1) != '/')//start of <label> { //string substr(int pos = 0,int n = npos) const;//返回pos开始的n个字符组成的字符串 bool isSub=false; string lableName = str.substr(sidx+1,eidx-sidx-1); if(label.Label != "") isSub = true;//start of subitem's <label> //find </label> int sidx2=0,eidx2=0; sidx2 = str.rfind(m_StartLabel); eidx2 = str.rfind(m_EndLabel); if(eidx != eidx2 && str.at(sidx2+1) == '/')//<label>content</label>, has </label> { string strelab = str.substr(sidx2+2,eidx2-sidx2-2); if(strelab == lableName){ if(isSub){ stMyDataLabel sublabel; sublabel.Label = lableName; sublabel.Content.clear(); sublabel.Content.push_back( str.substr(eidx+1,sidx2-eidx-1) ); label.SubItems.push_back(sublabel); } else{ label.Label = lableName; label.Content.clear(); label.Content.push_back( str.substr(eidx+1,sidx2-eidx-1) ); } continue; } else{ return false; } } else{ if(isSub){ stMyDataLabel sublabel; sublabel.Label = lableName; label.SubItems.push_back(sublabel); int curIdx = label.SubItems.size()-1; bool bres = ParseLabel(myin,label.SubItems.at(curIdx)); if(!bres) return false;//subitem format error else continue;// } else label.Label = lableName; } } else{// </lable> string elabel = str.substr(sidx+2,eidx-sidx-2); if(elabel == label.Label){//end of this label return true; } else{//format error return false; } } } else{//content if(label.Label == "data" || label.Label == "DATA"){//data block // 判断字符串是不是数字 stringstream sin(str); float val; if(!(sin >> val))//不是数字 label.Content.push_back(str); else label.Values.push_back(val); } else//header { int idx = str.find('='); if(idx >= 0){//key=value string strkey = str.substr(0,idx); string strval = str.substr(idx+1); label.KeyVal.insert(make_pair(strkey,strval)); } else{ label.Content.push_back(str); } } } }while(flag); } // 判断字符串是不是数字 bool isNum(string str) { stringstream sin(str); double d; char c; if(!(sin >> d)) return false; if (sin >> c) return false; return true; } }; /*file format2: #xtype //comment [condition] 项1=值1 项2=值2 [/condition] [mea] key3=value3 key4=value4 [/mea] [data] phi rcs ang 1 1 1 2 2 2 [/data] */ class MyDataParse2{//数据类型2的解析类 public: stMyDataLabel m_Data;//single root char m_StartLabel; char m_EndLabel; public: MyDataParse2(){ m_StartLabel = '['; m_EndLabel = ']'; } void Init(){ m_Data.Clear(); } void Print(ostream& myout){ myout<<m_Data; } bool Parse(char* filename){ if(filename == NULL) return false; ifstream myin(filename,ios::in); string str; myin>>str; if(str.at(0) == '#'){ m_Data.Label = str.substr(1); } bool flag = true; while(!myin.eof() && flag){ flag = ParseLabel(myin,m_Data); } return flag; } bool ParseLabel(istream& myin,stMyDataLabel& label){ string str; bool flag = true; do{ myin >> str; if(str.substr(0,2) == "//")//comment,skip continue; int sidx=0,eidx=0; sidx = str.find(m_StartLabel); eidx = str.find(m_EndLabel); if(sidx >= 0 && eidx >= 0 && eidx > sidx){//[lable] or [/label] if(str.at(sidx+1) != '/')//start of <label> { //string substr(int pos = 0,int n = npos) const;//返回pos开始的n个字符组成的字符串 bool isSub=false; string lableName = str.substr(sidx+1,eidx-sidx-1); if(label.Label != "") isSub = true;//start of subitem's [label] if(isSub){ stMyDataLabel sublabel; sublabel.Label = lableName; label.SubItems.push_back(sublabel); int curIdx = label.SubItems.size()-1; bool bres = ParseLabel(myin,label.SubItems.at(curIdx)); if(!bres) return false;//subitem format error else continue;// } else label.Label = lableName; } else{// </lable> string elabel = str.substr(sidx+2,eidx-sidx-2); if(elabel == label.Label){//end of this label return true; } else{//format error return false; } } } else{//content if(label.Label == "data" || label.Label == "DATA"){//data block // 判断字符串是不是数字 stringstream sin(str); float val; if(!(sin >> val))//不是数字 label.Content.push_back(str); else label.Values.push_back(val); } else//header { int idx = str.find('='); if(idx >= 0){//key=value string strkey = str.substr(0,idx); string strval = str.substr(idx+1); label.KeyVal.insert(make_pair(strkey,strval)); } else{ label.Content.push_back(str); } } } }while(!myin.eof()); } }; #endif
测试（命令行方式）
MyDataParse parse; parse.Init(); parse.Parse("test.txt");//待解析的数据文件 parse.Print(cout);
测试数据：
//test11.txt <head> <type>xtype</type> <condition> key1=value1 key2=value2 </condition> <mea> key3=value3 key4=value4 </mea> </head> <data> phi rcs ang 1 2 3 2 3 4 </data> <data> phi rcs ang 3 4 5 4 5 6 </data>
//test12.txt <head> <type>xtype</type> <condition> key1=value1 key2=value2 </condition> <group> <part1> key3=value3 key4=value4 </part1> <part2> key3=value3 key4=value4 </part2> </group> </head> <data> phi rcs ang 1 2 3 2 3 4 </data> <data> phi rcs ang 3 4 5 4 5 6 </data>
//test21.txt #xtype //comment [condition] 项1=值1 项2=值2 [/condition] [mea] key3=value3 key4=value4 [/mea] [data] phi rcs ang 1 1 1 2 2 2 [/data] [data] phi rcs ang 3 1 1 4 2 2 [/data]
//test22.txt #xtype //comment [condition] 项1=值1 项2=值2 [/condition] [group] [part1] key3=value3 [/part1] [part2] key4=value4 key5=value5 [/part2] [/group] [data] phi rcs ang 1 1 1 2 2 2 [/data] [data] phi rcs ang 3 1 1 4 2 2 [/data]