使用标准库用 strtok 替换循环 [英] Replace a loop with strtok by using Standard Library

问题描述

我有一个 strtok 解决的问题(从字符串中拆分子字符串)但是我意识到 strtok 是不安全的.我想使用 C++ 标准库的一些更现代的部分.

I have a problem that strtok solves (split substrings from a string) but I realize that strtok is not safe. I want to use some more modern parts of the C++ standard library.

我应该用什么代替?

static int ParseLine(std::string line, 
                     std::string seps, 
                     int startIdx, 
                     std::vector<CNode>& collection)
{
    if (startIdx > collection.size())
    {
        throw std::invalid_argument("the start index is out of range");
    }
    char buf[2000];
    strcpy_s(buf, line.c_str());
    auto idx = startIdx;
    for (auto objectType = strtok(buf, seps.c_str()); objectType != nullptr; idx++)
    {
        if (idx == collection.size())
        {
            collection.push_back(CNode(idx));
        }
        collection[idx].SetObjectType(objectType);
        objectType = strtok(nullptr, seps.c_str());
    }
    return (idx - 1);
}

这里是一个完整的示例，它与_CRT_SECURE_NO_WARNINGS:

Here a complete sample that compiles with _CRT_SECURE_NO_WARNINGS:

#include <string>
#include <vector>
#include <iostream>

class CObject
{
    std::string _objectType;

public:
                CObject() : _objectType("n/a") {}
    void        SetObjectType(std::string objectType) { _objectType = objectType; }
    std::string GetObjectType() const { return _objectType; }

};

class CNode
{
    int     _id;
    CObject _object;

public:
    explicit CNode(int id) : _id(id) {}
    void        SetObjectType(std::string objectType) { _object.SetObjectType(objectType); }
    std::string GetObjectType() const { return _object.GetObjectType(); }

};

// Update the collection of nodes with the object types found in the line specified.
// Return the number of elements in the collection.
static int ParseLine(std::string line, std::string seps, int startIdx, std::vector<CNode>& collection)
{
    if (startIdx > collection.size())
    {
        throw std::invalid_argument("the start index is out of range");
    }
    char buf[2000];
    strcpy_s(buf, line.c_str());
    auto idx = startIdx;
    for (auto objectType = strtok(buf, seps.c_str()); objectType != nullptr; idx++)
    {
        if (idx == collection.size())
        {
            collection.push_back(CNode(idx));
        }
        collection[idx].SetObjectType(objectType);
        objectType = strtok(nullptr, seps.c_str());
    }
    return (idx - 1);
}

int main()
{
    std::string seps = ".";
    //                   2   3   4   5   6   7   8   9
    std::string line = "abc.def.ghi.klm.nop.qrs.tuv.wxyz";
    std::vector<CNode> collection { CNode(0), CNode(1), CNode(2) , CNode(3) , CNode(4) , CNode(5) };
    auto startAt = 2;
    try
    {
        auto collection_size = ParseLine(line, seps, startAt, collection);
        std::cout << collection_size << std::endl;
        for (auto value : collection)
        {
            std::cout << value.GetObjectType() << std::endl;
        }
    }
    catch (std::invalid_argument& e)
    {
        std::cout << " out of range exception " << e.what() << std::endl;
    }
    return 0;
}

推荐答案

这里有两个分割分隔字符串的例子.

Here are two examples of splitting a delimited string.

第一个将 std::getline 与字符串流一起使用，指定分隔符而不是使用默认的换行符.此技术只能使用单字符分隔符.

The first uses std::getline with a string stream, specifying a separator character instead of using the default newline character. Only single-character separators may be used with this technique.

第二个示例使用了 库，它允许使用任意长度的分隔符，并且让您更好地控制如何识别分隔符.请注意，点字符必须在正则表达式规范中转义，因为在正则表达式语言中，."用作通配符.

The second example uses the <regex> library, which allows separators of arbitrary length and also gives you more control over how a separator is recognized. Note that the dot character must be escaped in the regex specification, because in the regex language, "." acts as a wildcard.

#include <iostream>
#include <sstream>
#include <vector>
#include <regex>

std::vector<std::string> GetlineSplit(std::string const& line) {
    static const char sep = '.';
    std::istringstream liness{line};
    std::vector<std::string> fields;
    for(std::string field; std::getline(liness, field, sep); ) {
        fields.push_back(field);
    }
    return fields;
}

std::vector<std::string> RegexSplit(std::string const& line) {
    std::regex seps("\\."); // the dot character needs to be escaped in a regex
    std::sregex_token_iterator rit(line.begin(), line.end(), seps, -1);
    return std::vector<std::string>(rit, std::sregex_token_iterator());
}

int main() {
    std::string line = "abc.def.ghi.klm.nop.qrs.tuv.wxyz";

    std::cout << "getline split result:\n";
    auto fields_getline = GetlineSplit(line);
    for(const auto& field : fields_getline) {
        std::cout << field << '\n';
    }

    std::cout << "\nregex split result:\n";
    auto fields_regex = RegexSplit(line);
    for(const auto& field : fields_regex) {
        std::cout << field << '\n';
    }
}

这篇关于使用标准库用 strtok 替换循环的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！