Finding all occurrences of a character in a string

I have comma delimited strings I need to pull values from. The problem is these strings will never be a fixed size. So I decided to iterate through the groups of commas and read what is in between. In order to do that I made a function that returns every occurrence's position in a sample string.

Is this a smart way to do it? Is this considered bad code?

#include <string>
#include <iostream>
#include <vector>
#include <Windows.h>

using namespace std;

vector<int> findLocation(string sample, char findIt);

int main()
{
    string test = "19,,112456开发者_如何转开发.0,a,34656";
    char findIt = ',';

    vector<int> results = findLocation(test,findIt);
    return 0;
}

vector<int> findLocation(string sample, char findIt)
{
    vector<int> characterLocations;
    for(int i =0; i < sample.size(); i++)
        if(sample[i] == findIt)
            characterLocations.push_back(sample[i]);

    return characterLocations;
}

vector<int> findLocation(string sample, char findIt)
{
    vector<int> characterLocations;
    for(int i =0; i < sample.size(); i++)
        if(sample[i] == findIt)
            characterLocations.push_back(sample[i]);

    return characterLocations;
}

As currently written, this will simply return a vector containing the int representations of the characters themselves, not their positions, which is what you really want, if I read your question correctly.

Replace this line:

characterLocations.push_back(sample[i]);

with this line:

characterLocations.push_back(i);

And that should give you the vector you want.

If I were reviewing this, I would see this and assume that what you're really trying to do is tokenize a string, and there's already good ways to do that.

Best way I've seen to do this is with boost::tokenizer. It lets you specify how the string is delimited and then gives you a nice iterator interface to iterate through each value.

using namespace boost;
string sample = "Hello,My,Name,Is,Doug";
escaped_list_seperator<char> sep("" /*escape char*/, ","/*seperator*/, "" /*quotes*/)

tokenizer<escaped_list_seperator<char> > myTokens(sample, sep)

//iterate through the contents
for (tokenizer<escaped_list_seperator<char>>::iterator iter = myTokens.begin();
     iter != myTokens.end();
     ++iter)
{
    std::cout << *iter << std::endl;
}

Output:

Hello
My
Name
Is
Doug

Edit If you don't want a dependency on boost, you can also use getline with an istringstream as in this answer. To copy somewhat from that answer:

std::string str = "Hello,My,Name,Is,Doug";
std::istringstream stream(str);
std::string tok1;

while (stream)
{
    std::getline(stream, tok1, ',');
    std::cout << tok1 << std::endl;
}

Output:

 Hello
 My
 Name
 Is
 Doug

This may not be directly what you're asking but I think it gets at your overall problem you're trying to solve.

Looks good to me too, one comment is with the naming of your variables and types. You call the vector you are going to return characterLocations which is of type int when really you are pushing back the character itself (which is type char) not its location. I am not sure what the greater application is for, but I think it would make more sense to pass back the locations. Or do a more cookie cutter string tokenize.

Well if your purpose is to find the indices of occurrences the following code will be more efficient as in c++ giving objects as parameters causes the objects to be copied which is insecure and also less efficient. Especially returning a vector is the worst possible practice in this case that's why giving it as a argument reference will be much better.

#include <string>
#include <iostream>
#include <vector>
#include <Windows.h>

using namespace std;

vector<int> findLocation(string sample, char findIt);

int main()
{

    string test = "19,,112456.0,a,34656";
    char findIt = ',';

    vector<int> results;
    findLocation(test,findIt, results);
    return 0;
}

void findLocation(const string& sample, const char findIt, vector<int>& resultList)
{
    const int sz = sample.size();

    for(int i =0; i < sz; i++)
    {
        if(sample[i] == findIt)
        {
            resultList.push_back(i);
        }
    }
}

How smart it is also depends on what you do with those subtstrings delimited with commas. In some cases it may be better (e.g. faster, with smaller memory requirements) to avoid searching and splitting and just parse and process the string at the same time, possibly using a state machine.