如何迭代由空格分隔的单词组成的字符串中的单词?

注意,我对C字符串函数或那种字符操作/访问不感兴趣。比起效率,我更喜欢优雅。我当前的解决方案:

#include <iostream>
#include <sstream>
#include <string>

using namespace std;

int main() {
    string s = "Somewhere down the road";
    istringstream iss(s);

    do {
        string subs;
        iss >> subs;
        cout << "Substring: " << subs << endl;
    } while (iss);
}

当前回答

这个呢

#include <string>
#include <vector>

using namespace std;

vector<string> split(string str, const char delim) {
    vector<string> v;
    string tmp;

    for(string::const_iterator i; i = str.begin(); i <= str.end(); ++i) {
        if(*i != delim && i != str.end()) {
            tmp += *i; 
        } else {
            v.push_back(tmp);
            tmp = ""; 
        }   
    }   

    return v;
}

其他回答

谢谢@Jairo Abdiel Toribio Cisneros。它对我有效,但您的函数返回一些空元素。因此,对于没有空的返回,我编辑了以下内容:

std::vector<std::string> split(std::string str, const char* delim) {
    std::vector<std::string> v;
    std::string tmp;

    for(std::string::const_iterator i = str.begin(); i <= str.end(); ++i) {
        if(*i != *delim && i != str.end()) {
            tmp += *i;
        } else {
            if (tmp.length() > 0) {
                v.push_back(tmp);
            }
            tmp = "";
        }
    }

    return v;
}

使用:

std::string s = "one:two::three";
std::string delim = ":";
std::vector<std::string> vv = split(s, delim.c_str());

在处理空格作为分隔符时,使用std::istream_iterator<T>的明显答案已经给出,并得到了很多支持。当然,元素可能不是用空格分隔,而是用一些分隔符分隔。我没有找到任何答案,只是重新定义了空格的含义,将其称为分隔符,然后使用常规方法。

要改变流对空白的看法,只需使用(std::istream::imbue())将流的std::locale更改为std::ctype<char>方面,并定义空白的含义(也可以对std:::ctype<wchar_t>进行更改,但实际上略有不同,因为std::actype<char>是表驱动的,而std::type<wchar_t>是由虚拟函数驱动的)。

#include <iostream>
#include <algorithm>
#include <iterator>
#include <sstream>
#include <locale>

struct whitespace_mask {
    std::ctype_base::mask mask_table[std::ctype<char>::table_size];
    whitespace_mask(std::string const& spaces) {
        std::ctype_base::mask* table = this->mask_table;
        std::ctype_base::mask const* tab
            = std::use_facet<std::ctype<char>>(std::locale()).table();
        for (std::size_t i(0); i != std::ctype<char>::table_size; ++i) {
            table[i] = tab[i] & ~std::ctype_base::space;
        }
        std::for_each(spaces.begin(), spaces.end(), [=](unsigned char c) {
            table[c] |= std::ctype_base::space;
        });
    }
};
class whitespace_facet
    : private whitespace_mask
    , public std::ctype<char> {
public:
    whitespace_facet(std::string const& spaces)
        : whitespace_mask(spaces)
        , std::ctype<char>(this->mask_table) {
    }
};

struct whitespace {
    std::string spaces;
    whitespace(std::string const& spaces): spaces(spaces) {}
};
std::istream& operator>>(std::istream& in, whitespace const& ws) {
    std::locale loc(in.getloc(), new whitespace_facet(ws.spaces));
    in.imbue(loc);
    return in;
}
// everything above would probably go into a utility library...

int main() {
    std::istringstream in("a, b, c, d, e");
    std::copy(std::istream_iterator<std::string>(in >> whitespace(", ")),
              std::istream_iterator<std::string>(),
              std::ostream_iterator<std::string>(std::cout, "\n"));

    std::istringstream pipes("a b c|  d |e     e");
    std::copy(std::istream_iterator<std::string>(pipes >> whitespace("|")),
              std::istream_iterator<std::string>(),
              std::ostream_iterator<std::string>(std::cout, "\n"));   
}

大部分代码用于打包提供软分隔符的通用工具:合并一行中的多个分隔符。无法生成空序列。当流中需要不同的分隔符时,您可能会使用共享流缓冲区使用不同的设置流:

void f(std::istream& in) {
    std::istream pipes(in.rdbuf());
    pipes >> whitespace("|");
    std::istream comma(in.rdbuf());
    comma >> whitespace(",");

    std::string s0, s1;
    if (pipes >> s0 >> std::ws   // read up to first pipe and ignore sequence of pipes
        && comma >> s1 >> std::ws) { // read up to first comma and ignore commas
        // ...
    }
}

没有Boost,没有字符串流,只有标准的C库与std::string和std::list:C库函数配合使用,便于分析,C++数据类型便于内存管理。

空白被认为是换行符、制表符和空格的任意组合。空白字符集由wschars变量建立。

#include <string>
#include <list>
#include <iostream>
#include <cstring>

using namespace std;

const char *wschars = "\t\n ";

list<string> split(const string &str)
{
  const char *cstr = str.c_str();
  list<string> out;

  while (*cstr) {                     // while remaining string not empty
    size_t toklen;
    cstr += strspn(cstr, wschars);    // skip leading whitespace
    toklen = strcspn(cstr, wschars);  // figure out token length
    if (toklen)                       // if we have a token, add to list
      out.push_back(string(cstr, toklen));
    cstr += toklen;                   // skip over token
  }

  // ran out of string; return list

  return out;
}

int main(int argc, char **argv)
{
  list<string> li = split(argv[1]);
  for (list<string>::iterator i = li.begin(); i != li.end(); i++)
    cout << "{" << *i << "}" << endl;
  return 0;
}

Run:

$ ./split ""
$ ./split "a"
{a}
$ ./split " a "
{a}
$ ./split " a b"
{a}
{b}
$ ./split " a b c"
{a}
{b}
{c}
$ ./split " a b c d  "
{a}
{b}
{c}
{d}

split的尾部递归版本(本身分裂为两个函数)。除了将字符串推入列表之外,所有对变量的破坏性操作都消失了!

void split_rec(const char *cstr, list<string> &li)
{
  if (*cstr) {
    const size_t leadsp = strspn(cstr, wschars);
    const size_t toklen = strcspn(cstr + leadsp, wschars);

    if (toklen)
      li.push_back(string(cstr + leadsp, toklen));

    split_rec(cstr + leadsp + toklen, li);
  }
}

list<string> split(const string &str)
{
  list<string> out;
  split_rec(str.c_str(), out);
  return out;
}

另一种灵活快速的方式

template<typename Operator>
void tokenize(Operator& op, const char* input, const char* delimiters) {
  const char* s = input;
  const char* e = s;
  while (*e != 0) {
    e = s;
    while (*e != 0 && strchr(delimiters, *e) == 0) ++e;
    if (e - s > 0) {
      op(s, e - s);
    }
    s = e + 1;
  }
}

要将其与字符串向量一起使用(编辑:由于有人指出不继承STL类…hrmf;):

template<class ContainerType>
class Appender {
public:
  Appender(ContainerType& container) : container_(container) {;}
  void operator() (const char* s, unsigned length) { 
    container_.push_back(std::string(s,length));
  }
private:
  ContainerType& container_;
};

std::vector<std::string> strVector;
Appender v(strVector);
tokenize(v, "A number of words to be tokenized", " \t");

就是这样!这只是使用tokenizer的一种方式,比如如何计数单词:

class WordCounter {
public:
  WordCounter() : noOfWords(0) {}
  void operator() (const char*, unsigned) {
    ++noOfWords;
  }
  unsigned noOfWords;
};

WordCounter wc;
tokenize(wc, "A number of words to be counted", " \t"); 
ASSERT( wc.noOfWords == 7 );

受限于想象力;)

我使用这个simpleton是因为我们得到了字符串类“特殊”(即非标准):

void splitString(const String &s, const String &delim, std::vector<String> &result) {
    const int l = delim.length();
    int f = 0;
    int i = s.indexOf(delim,f);
    while (i>=0) {
        String token( i-f > 0 ? s.substring(f,i-f) : "");
        result.push_back(token);
        f=i+l;
        i = s.indexOf(delim,f);
    }
    String token = s.substring(f);
    result.push_back(token);
}