如何有效地将值从std :: vector格式化为std :: string？

Question

I have a program that takes a std::vector<uint8_t> and returns a std::string formatted as hex characters followed by the ASCII text like so:

03 00 00 54 00 00 00 08 00 00 00 00 00 00 00 00 ASCII ...T............

74 21 B8 30 00 2C 2E 31 62 30 74 21 A8 30 00 2C ASCII t!.0.,.1b0t!.0.,

This is the main part of the code:

std::vector<uint8_t> value;
std::stringstream printed_line;
std::stringstream tempHexLine;
std::stringstream tempAsciiLine;

for (size_t printed_bytes = 0; printed_bytes < value.size(); printed_bytes += bytes_line) {

    for (int i = 0; i < bytes_line; ++i) {
        tempHexLine << std::setfill('0') << std::setw(2) << std::uppercase << std::hex << static_cast<uint16_t>(value[printed_bytes + i]) << " ";
    }

    for (int i = 0; i < bytes_line; ++i) {

            if(isprint(value[printed_bytes + i])) {

                if (value[printed_bytes + i] == 60) {
                    tempAsciiLine << "&lt;";

                } else if(value[printed_bytes + i] == 62) {
                    tempAsciiLine << "&gt;";

                } else {
                    tempAsciiLine << static_cast<char>(value[printed_bytes + i]);
                }

            } else {
                tempAsciiLine << ".";
            }
        }
    }

    printed_line << tempHexLine.str() << "   " << tempAsciiLine .str();

The problem that I'm trying to solve is that when the vector size is large (> 1000 elements), the this takes a long time - approximately 70% of the samples using Very Sleepy where to the ::operator<< function.

What is the fastest way to apply a format like this to the values in the vector? There are well defined chucks of data that the vector needs to be broken into, so it seems like outputting a byte at a time is an inefficient way.

Answer 1

This should be trivial to write efficiently. Try this:

#include <algorithm>
#include <cstdint>
#include <iterator>
#include <string>
#include <vector>

std::string to_hex(std::vector<uint8_t> const & v)
{
    std::string result;
    result.reserve(4 * v.size() + 6);

    for (uint8_t c : v)
    {
        static constexpr char alphabet[] = "0123456789ABCDEF";

        result.push_back(alphabet[c / 16]);
        result.push_back(alphabet[c % 16]);
        result.push_back(' ');
    }

    result.append("ASCII ", 6);
    std::copy(v.begin(), v.end(), std::back_inserter(result));

    return result;
}

Answer 2

You might use some conversion tables for individual characters:

#include <cctype>
#include <cstring>
#include <iostream>
#include <vector>

class HexadecimalTable
{
    public:
    static const unsigned ElementSize = 4;

    HexadecimalTable();

    const char* operator [] (std::uint8_t value) const {
        return m_table[value];
    }

    private:
    typedef char hex_type[ElementSize];
    hex_type m_table[256];
};

HexadecimalTable::HexadecimalTable() {
    static const char hex[] = "0123456789ABCDEF";
    for(int i = 0; i < 256; ++i) {
        m_table[i][0] = hex[i / 16];
        m_table[i][1] = hex[i % 16];
        m_table[i][2] = 0;
    }
}

class EntityTable
{
    public:
    static const unsigned ElementSize = 8;

    EntityTable();

    const char* operator [] (std::uint8_t value) const {
        return m_table[value];
    }

    private:
    typedef char entity_type[ElementSize];

    entity_type m_table[256];
};

EntityTable::EntityTable() {
    // Printable and non printable characters
    for(int i = 0; i < 256; ++i) {
        if(std::isprint(i)) m_table[i][0] = char(i);
        else m_table[i][0] = '.';
        m_table[i][1] = 0;
    }
    // Special characters
    std::strcpy(m_table['<'], "&lt;");
    std::strcpy(m_table['>'], "&gt;");
    // ...
}

std::string hex_display(std::vector<uint8_t> const & v) {
    HexadecimalTable hexadecimal;
    EntityTable entities;

    std::string result;
    result.reserve((HexadecimalTable::ElementSize + EntityTable::ElementSize) * v.size());

    for (uint8_t c : v) {
        result.append(hexadecimal[c]);
        result.push_back(' ');
    }

    for (uint8_t c : v)
        result.append(entities[c]);

    return result;
}

int main()
{
    std::vector<std::uint8_t> v = {
        0x74, 0x21, 0xB8, 0x30,
        0x00, 0x2C, 0x2E, 0x31,
        0x62, 0x30, 0x74, 0x21,
        0xA8, 0x30,   60,   62
    };
    std::cout << hex_display(v) << std::endl;
    return 0;
}