C ++-查找功能无法在子字符串上执行

Question

I do have got a problem with an algorithm, which I want to use to split command line into several substrings. Eg the string "Hello World -n filename" should be sliced into "Hello" "World" and "-n filename" .

Here is my whole code example:

string hello = "Hello World -n filename";
uint64_t startIndex = 0;
uint64_t endIndex = hello.length() - 1;

while(startIndex < endIndex) {
    uint64_t nextWhiteSpaceIndex;
    string value;

    if(hello.at(startIndex) != '-') {

        nextWhiteSpaceIndex = hello.substr(startIndex).find(" ");
        value = hello.substr(startIndex, nextWhiteSpaceIndex);
        cout << value << endl;

    } else {

        nextWhiteSpaceIndex = hello.substr(hello.substr(startIndex).find(" ")).find(" ");
        value = hello.substr(startIndex, nextWhiteSpaceIndex);
        cout << value << endl;

    }
    startIndex = nextWhiteSpaceIndex + 1;
}

And I do have problems with this command:

nextWhiteSpaceIndex = hello.substr(startIndex).find(" ");

This is placed within the while-loop and it seems like the...

.substr(startIndex)

... part is completely ignored. The first loop run works fine, but during the second/following the nextWhiteSpaceIndex does not get the right next index assigned. It always prints "Hello" "World" "World" "World" "World" and continues to print "World" .

Do you guys have a hint, why this does not work? I could not find an appropriate explanation during my research through the web.

Answer 1

Can't you do something like

#include <sstream>
#include <stdio>
#include <vector>

using namespace std;

int main ()
{
    string hello = "Hello World -n filename";

    stringstream ss (hello);
    vector<string> v;
    string s, t;

    while (ss >> s)
    {
        if (s[0] == '-')
        {   
            ss >> t;
            v.push_back (s + " " + t); 
        }   
        else
            v.push_back (s);
    }

    for (auto i : v)
        clog << i << endl;

    return 0;
}

produces

$ ./a.out
Hello
World
-n filename

Answer 2

If you output the value for nextWhiteSpaceIndex you will always see: 5, 5, 5, 5... It's an index relative to startIndex , so simply changing the last line into startIndex += nextWhiteSpaceIndex + 1; would probably quickfix the issue.

(Aren't you taking too many substrings? std::string::find can take the search start index as argument, so you can perform the whole search on always the same buffer.)

Answer 3

hello is never changed, yet you're using find on only part of it ( substr ), and then calling substr(startIndex) on the whole string ( hello ) over and over again.

1. "Hello World -n filename".find(" ") -> 5
2. "World -n filename".find(" ") -> 5 again ("World" is the same length as "Hello")
3. "World -n filename".find(" ") -> 5

You could use std::string::find 's second argument ( size_type pos ) to specify the start offset where to start the search and create fewer temporary strings:

#include <iostream>

using namespace std;

int main() {
  const string hello = "Hello World -n filename";

  size_t startIndex = 0, pos = 0;
  bool eat = true;

  while(true) {
    pos = hello.find('\x20', pos);

    if(pos == string::npos) {
      cout << hello.substr(startIndex) << endl;
      break;
    }
    else if(eat && hello[startIndex] == '-') {
      eat = false;
      ++pos;
      continue;
    }

    cout << hello.substr(startIndex, pos - startIndex) << endl;

    startIndex = ++pos;
    eat = true;
  }

  return 0;
}

Output:

$ c++ main.cpp && ./a.out
Hello
World
-n filename