Partners: KDAB Whole Tomato Software CppDepend

03 December 2018

How to Use The Newest C++ String Conversion Routines - std::from_chars

How To Use std::from_chars, C++17

With C++17 we get another facility to handle the conversion between text and numbers. Why should we care about the new routines? Are they better in any way?

Intro

C++, before C++17, offered several options when it comes to string conversion:

  • sprintf / snprintf
  • sscanf
  • atol
  • strtol
  • strstream
  • stringstream
  • to_string
  • stoi and similar functions

And with C++17 you get another option: std::from_chars! Wasn’t the old stuff good enough? Why do we need new methods?

In short: because from_chars is low-level, and offers the best possible performance.

The new conversion routines are:

  • non-throwing
  • non-allocating
  • no locale support
  • memory safety
  • error reporting gives additional information about the conversion outcome

The API might not be the most friendly to use, but it’s easy enough to wrap it into some facade.

A simple example:

const std::string str { "12345678901234" };
int value = 0;
std::from_chars(str.data(),str.data() + str.size(), value);
// error checking ommited...

The new methods are available in the following compilers

  • Visual Studio 2017 15.9 - full support (from_chars and also to_chars) (see notes about changes in 15.8 and in 15.9)
  • GCC - 8.0 - still in progress, only integer support
  • Clang 7.0 - still in progress, only integer support

If you want to read more about existing conversion routines, new ones and also see some benchmarks you can see two great posts at @fluentcpp:
How to Convert a String to an int in C++ and How to *Efficiently* Convert a String to an int in C++ written by JFT.

The Series

This article is part of my series about C++17 Library Utilities. Here’s the list of the topics in the series:

Resources about C++17 STL:

Let’s have a look at the API now.

Converting From Characters to Numbers: from_chars

std::from_chars is a set of overloaded functions: for integral types and floating point types.

For integral types we have the following functions:

std::from_chars_result from_chars(const char* first, 
                                  const char* last, 
                                  TYPE &value,
                                  int base = 10);

Where TYPE expands to all available signed and unsigned integer types and char.

base can be a number ranging from 2 to 36.

Then there’s the floating point version:

std::from_chars_result from_chars(const char* first, 
                   const char* last, 
                   FLOAT_TYPE& value,
                   std::chars_format fmt = std::chars_format::general);

FLOAT_TYPE expands to float, double or long double.

chars_format is an enum with the following values: scientific, fixed, hex and general (which is a composition of fixed and scientific).

The return value in all of those functions (for integers and floats) is from_chars_result:

struct from_chars_result {
    const char* ptr;
    std::errc ec;
};

from_chars_result holds valuable information about the conversion process.

Here’s the summary:

Return Condition State of from_chars_result
Success ptr points at the first character not matching the pattern, or has the value equal to last if all characters match and ec is value-initialized.
Invalid conversion ptr equals first and ec equals std::errc::invalid_argument. value is unmodified.
Out of range The number it too large to fit into the value type. ec equals std::errc::result_out_of_range and ptr points at the first character not matching the pattern. value is unmodified.

The new routines are very low level, so you might be wondering why is that. Titus Winters added a great summary in comments:

The intent of those APIs is *not* for people to use them directly, but to build more interesting/useful things on top of them. These are primitives, and we (the committee) believe they have to be in the standard because there isn't an efficient way to do these operations without calling back out to nul-terminated C routines

Examples

Here are two examples of how to convert a string into a number using from_chars, to int andfloat.

Integral types

#include <charconv> // from_char, to_char
#include <string>
#include <iostream>

int main()
{
    const std::string str { "12345678901234" };
    int value = 0;
    const auto res = std::from_chars(str.data(), 
                                     str.data() + str.size(), 
                                     value);

    if (res.ec == std::errc())
    {
        std::cout << "value: " << value 
                  << ", distance: " << res.ptr - str.data() << '\n';
    }
    else if (res.ec == std::errc::invalid_argument)
    {
        std::cout << "invalid argument!\n";
    }
    else if (res.ec == std::errc::result_out_of_range)
    {
        std::cout << "out of range! res.ptr distance: " 
                  << res.ptr - str.data() << '\n';
    }
}

The example is straightforward, it passes a string str into from_chars and then displays the result with additional information if possible.

Run the code below and change the str value to see the output:

Does "12345678901234" fit into the number? Or you see some errors from the conversion API?

Floating Point

To get the floating point test, we can replace the top lines of the previous example with:

// works with MSVC only so far (Dec 2018)
const std::string str { "16.78" };
double value = 0;
const auto format = std::chars_format::general;
const auto res = std::from_chars(str.data(), 
                                 str.data() + str.size(), 
                                 value, 
                                 format);

Here’s the example output that we can get:

str value format value output
1.01 fixed value: 1.01, distance 4
-67.90000 fixed value: -67.9, distance: 9
20.9 scientific invalid argument!, res.ptr distance: 0
20.9e+0 scientific value: 20.9, distance: 7
-20.9e+1 scientific value: -209, distance: 8
F.F hex value: 15.9375, distance: 3
-10.1 hex value: -16.0625, distance: 5

The general format is a combination of fixed and scientific so it handles regular floating point string with the additional support for e+num syntax.

A bit about Performance

I did some benchmark, and the new routines are blazing fast!

Some numbers:

  • On GCC it’s ~4,5x faster than stoi, 2,2x faster than atoi and almost 50x faster than istringstream.
  • On Clang it’s ~3,5x faster than stoi, 2.7x faster than atoi and 60x faster than istringstream!
  • MSVC performs ~3x faster than stoi, ~2x faster than atoi and almost 50x faster than istringstream

The whole benchmark might be however a story for another post, so I’m not sharing the code yet.

Summary

If you want to convert text into a number and you don’t need any extra stuff like locale support then std::from_chars might be the best choice. It offers great performance, and what’s more, you’ll get a lot of information about the conversion process (for example how much characters were scanned).

The routines might be especially handy with parsing JSON files, 3d textual model representation (like OBJ file formats), etc.

Your Turn

Have you played with the new conversion routines?
What do you usually use to convert text into numbers?

Get my free ebook about C++17!

More than 50 pages about the new Language Standard.

C++17 in detail, by Bartlomiej Filipek

© 2017, Bartlomiej Filipek, Blogger platform
Any opinions expressed herein are in no way representative of those of my employers.
This site contains ads or referral links, which provide me with a commission. Thank you for your understanding.