Converting String to Double -- Loses Precision - c++

I am having trouble converting a string to a double. I am given a string with lat/long coordinates in the format of 33.9425/N 118.4081/W
I first call my function trimLastChar(std::string& input) twice, which will remove the the North, South, East, West characters and then the forward slash. This function correctly returns 33.9425 and 118.4081 respectfully as a std::string.
I am using the following code to convert my std::string to a double...however the problem is, the conversion losses precision -- I suspect it gets rounded?.
// Location In String is what trimLastChar returns
std::stringstream stream(locationInString);
std::cout << "DEBUG: before: " << locationInString << " ";
// output is a double* output = new double passed by reference to my function
stream >> output;
std::cout << output << std::endl;
The output in this case would produce:
33.9425
118.408
As you notice, the correct value should be 118.4081 but the 1 is missing...
Any ideas how to fix? Or more important, why is this happening?

The precision wasn't lost on input. It's being lost on output.
#include <iostream>
using std::cout;
using std::endl;
int main()
{
double v = 118.4081;
cout << v << endl;
cout.precision(10);
cout << v << endl;
}
outputs:
$ g++ -Wall x.cpp && ./a.out
118.408
118.4081
$

You probably have more digits than the output shows. By default, only a small number of digits is shown, you need to use std::setprecision to see more digits. Try
std::cout << std::setprecision(10) << output << std::endl;

Related

Setprecision in a function is also applying in another function. I can't seem to know why [duplicate]

I want to control the precision for a double during a comparison, and then come back to default precision, with C++.
I intend to use setPrecision() to set precision. What is then syntax, if any, to set precision back to default?
I am doing something like this
std::setPrecision(math.log10(m_FTOL));
I do some stuff, and I would like to come back to default double comparison right afterwards.
I modified like this, and I still have some errors
std::streamsize prec = std::ios_base::precision();
std::setprecision(cmath::log10(m_FTOL));
with cmath false at compilation, and std::ios_base also false at compilation. Could you help?
You can get the precision before you change it, with std::ios_base::precision and then use that to change it back later.
You can see this in action with:
#include <ios>
#include <iostream>
#include <iomanip>
int main (void) {
double d = 3.141592653589;
std::streamsize ss = std::cout.precision();
std::cout << "Initial precision = " << ss << '\n';
std::cout << "Value = " << d << '\n';
std::cout.precision (10);
std::cout << "Longer value = " << d << '\n';
std::cout.precision (ss);
std::cout << "Original value = " << d << '\n';
std::cout << "Longer and original value = "
<< std::setprecision(10) << d << ' '
<< std::setprecision(ss) << d << '\n';
std::cout << "Original value = " << d << '\n';
return 0;
}
which outputs:
Initial precision = 6
Value = 3.14159
Longer value = 3.141592654
Original value = 3.14159
Longer and original value = 3.141592654 3.14159
Original value = 3.14159
The code above shows two ways of setting the precision, first by calling std::cout.precision (N) and second by using a stream manipulator std::setprecision(N).
But you need to keep in mind that the precision is for outputting values via streams, it does not directly affect comparisons of the values themselves with code like:
if (val1== val2) ...
In other words, even though the output may be 3.14159, the value itself is still the full 3.141592653590 (subject to normal floating point limitations, of course).
If you want to do that, you'll need to check if it's close enough rather than equal, with code such as:
if ((fabs (val1 - val2) < 0.0001) ...
Use C++20 std::format and {:.2} instead of std::setprecision
Finally, this will be the superior choice once you can use it:
#include <format>
#include <string>
int main() {
std::cout << std::format("{:.3} {:.4}\n", 3.1415, 3.1415);
}
Expected output:
3.14 3.145
This will therefore completely overcome the madness of modifying std::cout state.
The existing fmt library implements it for before it gets official support: https://github.com/fmtlib/fmt Install on Ubuntu 22.04:
sudo apt install libfmt-dev
Modify source to replace:
<format> with <fmt/core.h>
std::format to fmt::format
main.cpp
#include <iostream>
#include <fmt/core.h>
int main() {
std::cout << fmt::format("{:.3} {:.4}\n", 3.1415, 3.1415);
}
and compile and run with:
g++ -std=c++11 -o main.out main.cpp -lfmt
./main.out
Output:
3.14 3.142
See also:
How do I print a double value with full precision using cout?
std::string formatting like sprintf
Pre C++20/fmt::: Save the entire state with std::ios::copyfmt
You might also want to restore the entire previous state with std::ios::copyfmt in these situations, as explained at: Restore the state of std::cout after manipulating it
main.cpp
#include <iomanip>
#include <iostream>
int main() {
constexpr float pi = 3.14159265359;
constexpr float e = 2.71828182846;
// Sanity check default print.
std::cout << "default" << std::endl;
std::cout << pi << std::endl;
std::cout << e << std::endl;
std::cout << std::endl;
// Change precision format to scientific,
// and restore default afterwards.
std::cout << "modified" << std::endl;
std::ios cout_state(nullptr);
cout_state.copyfmt(std::cout);
std::cout << std::setprecision(2);
std::cout << std::scientific;
std::cout << pi << std::endl;
std::cout << e << std::endl;
std::cout.copyfmt(cout_state);
std::cout << std::endl;
// Check that cout state was restored.
std::cout << "restored" << std::endl;
std::cout << pi << std::endl;
std::cout << e << std::endl;
std::cout << std::endl;
}
GitHub upstream.
Compile and run:
g++ -ggdb3 -O0 -std=c++11 -Wall -Wextra -pedantic -o main.out main.cpp
./main.out
Output:
default
3.14159
2.71828
modified
3.14e+00
2.72e+00
restored
3.14159
2.71828
Tested on Ubuntu 19.04, GCC 8.3.0.
You need to keep track of your current precison and then reset back to the same once done with your operations with required modified precison. For this you can use std::ios_base::precision:
streamsize precision ( ) const;
streamsize precision ( streamsize prec );
The first syntax returns the value of the current floating-point precision field for the stream.
The second syntax also sets it to a new value.
setprecision() can be used only for output operations and cannot be used for comparisons
To compare floats say a and b , you have to do it explicitly like this:
if( abs(a-b) < 1e-6) {
}
else {
}
You can use cout << setprecision(-1)

Conversion of string to double without scientific notation in C++

I have a string variable and want to convert it into double without any scientific notation. I tried using std::stod but that doesn't work.
std::stringstream timestamp;
timestamp << t_val;
string test = timestamp.str();
cout << test << endl; // this gives 1506836639.96
double d = std::stod(test);
cout << d << endl; // This gives 1.50684e+09 instead of 1506836639.96
I tried using setprecision and fixed but I couldn't store the result into a variable. Is there a way I can store the value of test (1506836639.96) as a double?
The scientific notation has to do with std::cout, not the way the value is stored, so you must use std::fixed before you print the value:
std::cout << std::fixed << std::setprecision(2) << d << std::endl;
As you can see in the demo, this works fine, it should work for you as well.
As #goodvibration commented std::to_string also works but it's not possible to redefine, in a simple manner, the default number or decimal places in this case.
std::cout << std::to_string(d) << std::endl;
Live demo

Stop the rounding on float c++

I am getting an issue when trying to output my float using std::cout <<
I have the following values:
vector2f = {-32.00234098f, 96.129380f} //takes 2 floats (x, y)
output: -32.0023:96.1294
What I am looking for is:
output: -32.00234098:96.129380
The actual numbers could be vary from the 7 decimal places (.0000007) to 3 decimal places (.003) so setting a fixed rounding number does not work in this case.
Any help would be great as I have tried changed to doubles as well but to no avail.
Thanks in advance!
There are 2 problems.
you need to include <iomanip> and use the std::setprecision manipulator.
To get the level of accuracy you want you will need to use doubles rather than floats.
e.g.:
#include <iostream>
#include <iomanip>
int main()
{
auto x = -32.00234098f, y = 96.129380f;
std::cout << std::setprecision(8) << std::fixed << x << ":" << y << std::endl;
// doubles
auto a = -32.00234098, b = 96.129380;
std::cout << std::setprecision(8) << std::fixed << a << ":" << b << std::endl;
}
example output:
-32.00234222:96.12937927
-32.00234098:96.12938000
You can set the output precision of the stream using std::precision manipulator.
To print trailing zeroes up to the given precision like in your example output, you need to use std::fixed manipulator.

C++ precision - behaviour of setprecision

As I understand it the setprecision function specifies the minimal precision but when I run the following code I get only 3 numbers after the decimal point:
int main()
{
double a = 123.4567890;
double b = 123.4000000;
std::cout << std::setprecision(5) << a << std::endl; // Udesireble
std::cout.setf(std::ios::fixed);
std::cout << std::setprecision(5) << a << std::endl; // Desireble
std::cout << std::setprecision(5) << b << std::endl; // Udesireble
std::cout.unsetf(std::ios::fixed);
std::cout << std::setprecision(5) << b << std::endl; // Desireble
return 0;
}
which prints:
123.46 // Udesireble
123.45679 // Desireble
123.40000 // Udesireble
123.4 // Desireble
Is there any way I can avoid checking the number of digits after the decimal point myself in order to know whether to set fixed ?
My impression is that you will need to format to string first, and then replace trailing zeros with spaces.
For the streams, you can use two functions.
setfill(char_type c), which set the character to write, to match with the number of needed character (more information here)
There is the setw(int) function, which set the width of field of the value to display. (documentation here )
Using these functions, you may have a solution

Set back default floating point print precision in C++

I want to control the precision for a double during a comparison, and then come back to default precision, with C++.
I intend to use setPrecision() to set precision. What is then syntax, if any, to set precision back to default?
I am doing something like this
std::setPrecision(math.log10(m_FTOL));
I do some stuff, and I would like to come back to default double comparison right afterwards.
I modified like this, and I still have some errors
std::streamsize prec = std::ios_base::precision();
std::setprecision(cmath::log10(m_FTOL));
with cmath false at compilation, and std::ios_base also false at compilation. Could you help?
You can get the precision before you change it, with std::ios_base::precision and then use that to change it back later.
You can see this in action with:
#include <ios>
#include <iostream>
#include <iomanip>
int main (void) {
double d = 3.141592653589;
std::streamsize ss = std::cout.precision();
std::cout << "Initial precision = " << ss << '\n';
std::cout << "Value = " << d << '\n';
std::cout.precision (10);
std::cout << "Longer value = " << d << '\n';
std::cout.precision (ss);
std::cout << "Original value = " << d << '\n';
std::cout << "Longer and original value = "
<< std::setprecision(10) << d << ' '
<< std::setprecision(ss) << d << '\n';
std::cout << "Original value = " << d << '\n';
return 0;
}
which outputs:
Initial precision = 6
Value = 3.14159
Longer value = 3.141592654
Original value = 3.14159
Longer and original value = 3.141592654 3.14159
Original value = 3.14159
The code above shows two ways of setting the precision, first by calling std::cout.precision (N) and second by using a stream manipulator std::setprecision(N).
But you need to keep in mind that the precision is for outputting values via streams, it does not directly affect comparisons of the values themselves with code like:
if (val1== val2) ...
In other words, even though the output may be 3.14159, the value itself is still the full 3.141592653590 (subject to normal floating point limitations, of course).
If you want to do that, you'll need to check if it's close enough rather than equal, with code such as:
if ((fabs (val1 - val2) < 0.0001) ...
Use C++20 std::format and {:.2} instead of std::setprecision
Finally, this will be the superior choice once you can use it:
#include <format>
#include <string>
int main() {
std::cout << std::format("{:.3} {:.4}\n", 3.1415, 3.1415);
}
Expected output:
3.14 3.145
This will therefore completely overcome the madness of modifying std::cout state.
The existing fmt library implements it for before it gets official support: https://github.com/fmtlib/fmt Install on Ubuntu 22.04:
sudo apt install libfmt-dev
Modify source to replace:
<format> with <fmt/core.h>
std::format to fmt::format
main.cpp
#include <iostream>
#include <fmt/core.h>
int main() {
std::cout << fmt::format("{:.3} {:.4}\n", 3.1415, 3.1415);
}
and compile and run with:
g++ -std=c++11 -o main.out main.cpp -lfmt
./main.out
Output:
3.14 3.142
See also:
How do I print a double value with full precision using cout?
std::string formatting like sprintf
Pre C++20/fmt::: Save the entire state with std::ios::copyfmt
You might also want to restore the entire previous state with std::ios::copyfmt in these situations, as explained at: Restore the state of std::cout after manipulating it
main.cpp
#include <iomanip>
#include <iostream>
int main() {
constexpr float pi = 3.14159265359;
constexpr float e = 2.71828182846;
// Sanity check default print.
std::cout << "default" << std::endl;
std::cout << pi << std::endl;
std::cout << e << std::endl;
std::cout << std::endl;
// Change precision format to scientific,
// and restore default afterwards.
std::cout << "modified" << std::endl;
std::ios cout_state(nullptr);
cout_state.copyfmt(std::cout);
std::cout << std::setprecision(2);
std::cout << std::scientific;
std::cout << pi << std::endl;
std::cout << e << std::endl;
std::cout.copyfmt(cout_state);
std::cout << std::endl;
// Check that cout state was restored.
std::cout << "restored" << std::endl;
std::cout << pi << std::endl;
std::cout << e << std::endl;
std::cout << std::endl;
}
GitHub upstream.
Compile and run:
g++ -ggdb3 -O0 -std=c++11 -Wall -Wextra -pedantic -o main.out main.cpp
./main.out
Output:
default
3.14159
2.71828
modified
3.14e+00
2.72e+00
restored
3.14159
2.71828
Tested on Ubuntu 19.04, GCC 8.3.0.
You need to keep track of your current precison and then reset back to the same once done with your operations with required modified precison. For this you can use std::ios_base::precision:
streamsize precision ( ) const;
streamsize precision ( streamsize prec );
The first syntax returns the value of the current floating-point precision field for the stream.
The second syntax also sets it to a new value.
setprecision() can be used only for output operations and cannot be used for comparisons
To compare floats say a and b , you have to do it explicitly like this:
if( abs(a-b) < 1e-6) {
}
else {
}
You can use cout << setprecision(-1)