String reversal in C++

Question

I am trying to reverse the order of words in a sentence by maintaining the spaces as below.

[this is my test    string] ==> [string test my is    this]

I did in a step by step manner as,

[this is my test    string] - input string
[gnirts    tset ym si siht] - reverse the whole string - in-place
[string    test my is this] - reverse the words of the string - in-place
[string test my is    this] - string-2 with spaces rearranged

Is there any other method to do this ? Is it also possible to do the last step in-place ?

Answer 1

Your approach is fine. But alternatively you can also do:

• Keep scanning the input for words and spaces
• If you find a word push it onto stack S
• If you find space(s) enqueue the number of spaces into a queue Q

After this is done there will be N words on the stack and N-1 numbers in the queue.

While stack not empty do
 print S.pop
 if stack is empty break
 print Q.deque number of spaces
end-while

Answer 2

Here's an approach.

In short, build two lists of tokens you find: one for words, and another for spaces. Then piece together a new string, with the words in reverse order and the spaces in forward order.

#include <iostream>
#include <algorithm>
#include <vector>
#include <string>
#include <sstream>
using namespace std;

string test_string = "this is my test    string";

int main()
{
    // Create 2 vectors of strings.  One for words, another for spaces.
    typedef vector<string> strings;
    strings words, spaces;
    // Walk through the input string, and find individual tokens.
    // A token is either a word or a contigious string of spaces.
    for( string::size_type pos = 0; pos != string::npos; )
    {
        // is this a word token or a space token?
        bool is_char = test_string[pos] != ' ';
        string::size_type pos_end_token = string::npos;

        // find the one-past-the-end index for the end of this token
        if( is_char )
            pos_end_token = test_string.find(' ', pos);
        else
            pos_end_token = test_string.find_first_not_of(' ', pos);

        // pull out this token
        string token = test_string.substr(pos, pos_end_token == string::npos ? string::npos : pos_end_token-pos);
        // if the token is a word, save it to the list of words.
        // if it's a space, save it to the list of spaces
        if( is_char )
            words.push_back(token);
        else
            spaces.push_back(token);
        // move on to the next token
        pos = pos_end_token;
    }

    // construct the new string using stringstream
    stringstream ss;
    // walk through both the list of spaces and the list of words,
    // keeping in mind that there may be more words than spaces, or vice versa
    // construct the new string by first copying the word, then the spaces
    strings::const_reverse_iterator it_w = words.rbegin();
    strings::const_iterator it_s = spaces.begin();
    while( it_w != words.rend() || it_s != spaces.end() )
    {
        if( it_w != words.rend() )
            ss << *it_w++;
        if( it_s != spaces.end() )
            ss << *it_s++;
    }

    // pull a `string` out of the results & dump it
    string reversed = ss.str();
    cout << "Input: '" << test_string << "'" << endl << "Output: '" << reversed << "'" << endl;

}

Answer 3

I would rephrase the problem this way:

• Non-space tokens are reversed, but preserves their original order
  • The 5 non-space tokens ‘this’, ‘is’, ‘my’, ‘test’, ‘string’ gets reversed to ‘string’, ‘test’, ‘my’, ‘is’, ‘this’.
• Space tokens remain in the original order
  • The space tokens ‘ ‘, ‘ ‘, ‘ ‘, ‘ ‘ remains in original order between the new order of non-space tokens.

Following is a O(N) solution [N being the length of char array]. Unfortunately, it is not in place as OP wanted, but it does not use additional stack or queue either -- it uses a separate character array as a working space.

Here is a C-ish pseudo code.

work_array = char array with size of input_array
dst = &work_array[ 0 ]

for( i = 1; ; i++) {
   detect i’th non-space token in input_array starting from the back side
   if no such token {
      break;
   }
   copy the token starting at dst
   advance dst by token_size
   detect i’th space-token in input_array starting from the front side
   copy the token starting at dst
   advance dst by token_size
}

// at this point work_array contains the desired output,
// it can be copied back to input_array and destroyed