Parse string of integers into std vector?

Question

Possible Duplicate:
Parsing a comma-delimited std::string

I want to parse a string into an integer vector:

string s = "1;2;4;8;16;";
vector<int> n = parse_string(s);
// n == [1, 2, 4, 8, 16];

Of course, I can write a simple code with strtok and atoi. But, what would be a much shorter code with C++ boost? I never tried with Boost, but heard that it could simply your code pretty much as if using Python.

I think this question has been asked a few times already: http://stackoverflow.com/questions/1894886/parsing-a-comma-delimited-stdstring http://stackoverflow.com/questions/4328685/input-line-by-line-from-an-input-file-and-tokenize-using-strtok-and-the-output http://stackoverflow.com/questions/536148/c-string-parsing-python-style http://stackoverflow.com/questions/1511029/c-tokenize-a-string-and-include-delimiters http://stackoverflow.com/questions/3162108/a-better-way-to-split-a-string-into-an-array-of-strings-in-c-c-using-whitespac — Jared Krumsie, Dec 10 '11 at 04:29

Hauleth · Answer 1 · 2011-12-09T20:48:47.060

3

It can be done without Boost:

string s = "1;2;4;8;16";
vector<int> n;
transform(s.begin(), s.end(), [](char c){return c == ';' ? ' ' : c});
stringstream ss(s);
copy(istream_iterator<int>(ss), istream_iterator<int>(), back_inserter(n));

EDIT: If you wanna use only C++03 code you must write:

char semicolon_to_space(char c){
  return c == ';' ? ' ' : c
};

// ...

string s = "1;2;4;8;16";
vector<int> n;
transform(s.begin(), s.end(), semicolon_to_space);
stringstream ss(s);
copy(istream_iterator<int>(ss), istream_iterator<int>(), back_inserter(n))

edited Dec 09 '11 at 20:48

answered Dec 09 '11 at 20:29

Hauleth

22,873
4
61
112

It should be noted that this code is C++11 – John Dibling Dec 09 '11 at 20:39
It can be very easy rewrote to C++03. U need only to kick out lambda as external named function. – Hauleth Dec 09 '11 at 20:46
Looping over to replace the ';' with ' ' is a bit wasteful. You can do it in a single operation by adding a converter object to the istream_iterator see below. – Martin York Dec 09 '11 at 21:56
3

In pure C++03 (or in C++11) you can also use `std::replace` for the first step: `std::replace( s.begin(), s.end(), ';', ' ' );` which is also more readable than the lambda or the external function. – David Rodríguez - dribeas Dec 09 '11 at 23:20
first, for this you have to get either nonconst string or make it's copy – Alek86 Dec 11 '11 at 03:13
second, try to parse "1;2;3;4;5 6;7;8;9;0" in this function – Alek86 Dec 11 '11 at 03:13
I suppose that input is valid. I showed only way to do this not to check that input is correct (as he/she asked). – Hauleth Apr 21 '12 at 14:13

score 1 · Answer 2 · answered Dec 09 '11 at 20:37

1

I'm not a boost profi, so my code is not ideal:

#include <boost/algorithm/string/split.hpp>
#include <boost/algorithm/string/classification.hpp>
#include <boost/foreach.hpp>
#include <iostream>
#include <vector>
#include <string>

int main() 
{
    using namespace boost::algorithm;

    std::vector< std::string > result;
    split(result, "1;2;4;8;16;", is_any_of(";"));

    BOOST_FOREACH(const std::string& item, result)
    {
        std::cout << item << std::endl;
    }
}

you can dispose of is_any_of for sure

answered Dec 09 '11 at 20:37

Alek86

1,489
3
17
26

split is in namespace `boost`, not `boost::algorithm`. – Benjamin Lindley Dec 09 '11 at 20:52
in the file (boost/algorithm/string/split.hpp) it's in boost::algorithm – Alek86 Dec 09 '11 at 21:17
I see. Apparently, it's in both. – Benjamin Lindley Dec 09 '11 at 21:27

score 0 · Answer 3 · answered Dec 09 '11 at 20:31

I think that there are 2 separate algorithms should be used in this case. The first one will be a parser which will separate your values boost::split http://www.boost.org/doc/libs/1_48_0/doc/html/string_algo/reference.html#header.boost.algorithm.string.split_hpp should help. The second one will be the lexical cast which will convert your string integers to int values take a look to boost::lexical_cast<>.

score -1 · Answer 4 · answered Dec 09 '11 at 21:11

-1

uhm boost::tokenizer? Wondering if I'm missing something, but isn't that ideal for this application?

Example on http://www.boost.org/doc/libs/1_43_0/libs/tokenizer/char_separator.htm < that page can easily be modified to the needs here.

answered Dec 09 '11 at 21:11

paul23

8,799
12
66
149

Martin York · Answer 5 · 2011-12-10T16:40:03.757

-1

Its easier if they are space separated:
If that is a requirement that is easily added.

#include <vector>
#include <iterator>
#include <algorithm>

int main()
{
    string s = "1 2 4 8 16";
    std::stringstream s_stream(s);

    vector<int> n;
    std::copy(std::istream_iterator<int>(s_stream), std::istream_iterator<int>(),
              std::back_inserter(n)
             );
}

If you must have the ';' then a simple converter class can be inserted.

struct IntReader
{
    int value;
    operator int() {return value;}
    friend std::istream& operator>>(std::istream& stream, IntReader& data)
    {
        char x = 0;
        if ((stream >> data.value >> x) && (x != ';'))
        {   stream.setstate(std::ios::failbit);
        }
        return stream;
    }
};

Then just change the copy to use it:

    std::copy(std::istream_iterator<IntReader>(s_stream), std::istream_iterator<IntReader>(),
              std::back_inserter(n)
             );

edited Dec 10 '11 at 16:40

answered Dec 09 '11 at 21:30

Martin York

257,169
86
333
562

I have added initialization of the `x` variable, I hope you don't mind. Additionally, I am not sure of the implementation of the `operator>>`. If `(stream >> data.value >>x)`, then the `stream` should never have the `failbit` set. Maybe the condition should be negated? – David Rodríguez - dribeas Dec 09 '11 at 23:24
@DavidRodríguez-dribeas: I don't think initialization of x is necassery. In `stream >> data.value >> x` If any part fails then then failbit is already set and I do not need to do anything. If it work then we need to check the value of x. But since it worked we know it has been initialized and thus the check for `x != ';'` is valid. Thus if the read worked and x is not ';' then we set the fail bit. – Martin York Dec 10 '11 at 05:07
Then the `if` reads: *if we have correctly read the number and another character that is not `;` then clear the failbit*? If we have read correctly (necessary by the first half of the condition, then `failbit` is not set and thus the body of the `if` is useless, right? My problem is that I am not understanding the intent, should it be a `setstate` rather than a `clear`? *If we read the number and a character, but the character is not a `;` then something went wrong?* – David Rodríguez - dribeas Dec 10 '11 at 08:33
It's supposed to set the failbbit. My brain is obviously not working. Thanks. – Martin York Dec 10 '11 at 16:39

Parse string of integers into std vector?

5 Answers5