Identify a specific word in a C++ String, is there a function ready for that? Or just in the nail itself?

Question

Hello,

I’m getting a series of strings(lines) from a . txt file and putting them into a vector. But there’s a problem, strings come with an unwanted start, example:

  Depende: lsb-release
  Depende: libatk1.0-0
  Depende: libc6
  Depende: libcairo-gobject2
  Depende: libcairo2
  Depende: libdbus-1-3
  Depende: libdbus-glib-1-2
  Depende: libfontconfig1
  Depende: libfreetype6
  Depende: libgcc1
  Depende: libgdk-pixbuf2.0-0
  Depende: libglib2.0-0
  Depende: libgtk-3-0
  Depende: libpango-1.0-0
  Depende: libpangocairo-1.0-0
  Depende: libstartup-notification0
  Depende: libstdc++6
  Depende: libx11-6

I would like to remove the beginning of each of them, the "Depends". Is there a function ready for this or will I have to do it in the nail? Each line of this I’ve put inside a vector.

Answer 1

There are some ways to do this, one of them being with std::string::find_last_of and std::string::substr: finding the position of space after Depende: and creating a substring that starts from that position.

#include <string>
#include <cassert>

int main()
{
    std::string s = " Depende: lsb-release";
    const size_t p = s.find_last_of(' ');
    std::string pacote = s.substr(p + 1);
    assert(pacote == "lsb-release");
}

Another possible way is to use std::string::substr directly, passing the number of characters to be skipped, in case you know that the string will always start with " Depende: ".

Yet another way is to use std::string::replace, to locally remove this prefix from the string:

#include <string>
#include <cassert>

int main()
{
    std::string s = " Depende: lsb-release";
    const size_t p = s.find_last_of(' ');
    s.replace(0, p + 1, ""); //< Substitui " Depende: " com "".
    assert(s == "lsb-release");
}

There is also the solution with regex (as seen in Reply by @Marcosbanik), but I avoid using it because of its controversy.

Answer 2

regex_replace may be what you seek.

#include <iostream>
#include <regex>
#include <string>

int main() {
  std::string line1{"Depende: lsb-release"};
  std::string line2{"Independe: lsb-release"};
  std::regex exp("^Depende: ");
  std::cout << std::regex_replace(line1, exp, "") << "\n";
  std::cout << std::regex_replace(line2, exp, "") << "\n";
}

Answer 3

The answers were great, maybe I redo my code. I ended up finishing it on the nail. But I liked knowing that I can "automate" this. I solved it this way:

#ifndef PACOTE_H
#define PACOTE_H
#include <vector>
#include <fstream>
#include <string>

class Pacote
{
    public:
        /*Metodo que retorna o nome do pacote definido pelo usuário*/
        std::string GetnomeDoPacote(){
        return nomeDoPacote;}

        /*Metodo que define o nome do pacote a ser trabalhado*/
        void SetnomeDoPacote(std::string variavelNomeDoPacote){
        nomeDoPacote = variavelNomeDoPacote;}

        /*Metodo que retorna as dependencias do pacote*/
        std::vector<std::string> GetdependenciasDoPacote(){
        return dependenciasDoPacote;}

        /*Metodo que extrai as dependencias do pacote de um arquivo .txt linha a linha, para que elas sejam trabalhadas*/
        void SetdependenciasDoPacote(std::string TXT){
          std::string linha;
          std::ifstream dependenciasTXT (TXT);
          if (dependenciasTXT.is_open())
          {
            while ( getline(dependenciasTXT,linha) )
            {
              dependenciasDoPacote.push_back(linha);
            }
            /*O primeiro e o último elemento do arquivo .txt não nos enteressa, pois eles não contem o nome de
            nenhuma dependencia a ser baixada*/
            dependenciasDoPacote.erase(dependenciasDoPacote.begin());
            dependenciasDoPacote.pop_back();
            /*Para cada elemento adquirido através do comando 'apt-cache depends', temos as classificaçoes "Depende", "Recomendado" e "Sugere".
            Precisamos remover essas strings do vector que armazenará essas dependências, logo, o esquema abaixo busca pelo primeiro elemento dessas
            strings, afim de identifica-las e removêlas atráves do comando erase.*/
            for(int linha = 0; linha < dependenciasDoPacote.size(); linha++){
                for(int elementoDaLinha = 2; elementoDaLinha < sizeof(dependenciasDoPacote[linha]); elementoDaLinha++){
                    if(dependenciasDoPacote[linha][elementoDaLinha] == 'D'){
                        dependenciasDoPacote[linha].erase(dependenciasDoPacote[linha].begin(), dependenciasDoPacote[linha].begin()+10);
                    }
                    else
                    if(dependenciasDoPacote[linha][elementoDaLinha] == 'R'){
                        dependenciasDoPacote[linha].erase(dependenciasDoPacote[linha].begin(), dependenciasDoPacote[linha].begin()+12);
                    }
                    else
                    if(dependenciasDoPacote[linha][elementoDaLinha] == 'S'){
                        dependenciasDoPacote[linha].erase(dependenciasDoPacote[linha].begin(), dependenciasDoPacote[linha].begin()+9);
                    }
                }
            }
            dependenciasTXT.close();
          }

          else{
            dependenciasDoPacote.push_back("Não foi possível encontrar o arquivo de dependências! :(");
            }
        }

        /*Metodo que retorna o nome do repositório PPA do pacote, definido pelo usuario*/
        std::string GetrepositorioPPADoPacote(){
        return repositorioPPADoPacote;}

        /*Metodo que define o nome do repositório PPA a ser trabalhado*/
        void SetrepositorioPPADoPacote(std::string nomeDoRepositorioPPA){
        repositorioPPADoPacote = nomeDoRepositorioPPA; }

    private:
        std::string nomeDoPacote;
        std::vector<std::string> dependenciasDoPacote;
        std::string repositorioPPADoPacote;
};

#endif PACOTE_H

I used the Rase and Begin method. Okay, it wasn’t exactly on the nail, but it was almost on the nail. dependencesDoPackage[line]. Rase(dependencesDoPackage[line]. Begin(), dependencesDoPackage[line]. Begin()+9);

With Rase I define that I want to erase, and with Begin I say that I want to erase by the front. I’m working with a vector. the vector line 0 has a certain amount of elements, and if I say 'dependencesDoPackage[line]. Begin(), I refer to the first element(element 0). In the above passage, I say I want to delete from element 0 to element 9.

dependencesDoPackage[line]. Rase(dependencesDoPackage[line]. Begin() /of element 0/, dependencesDoPackage[line]. Begin()+9 /to element 9/);

This works because the . txt file being worked on follows a pattern. It always starts at the [line][2] element of the vector. And it always starts either with Recommended, or with Depends, or with Suggests. So, I know, through ifs and elses what I’m deleting or not. If recommended, starting with the element [line][2] by the letter 'R', I delete until such element, if it is Suggest, starting with the letter’S', I delete until such element.... Anyway. Please point out errors or bad practices in my code.