setprecision support

[young66]

What is the issue you have?

i try to run the follow code

#include <iostream>
#include <iomanip>
json testj;
testj["data1"] = 7.0000000001;//copy from var like 7.0
//some more data in json
testj["data2"] = 0.9999999999;//copy from var like 1.0
cout << setprecision(2);
cout << fixed;
cout << "precision=2 fix :" << 1.1111111111111 << ", json: " << testj<< endl;

Please describe the steps to reproduce the issue.

run the code

Can you provide a small but working code example?

yes ,see the sample

#include <iostream>
#include <iomanip>
json testj;
testj["data1"] = 7.0000000001;//copy from var with 7.0
//some more data in json
testj["data2"] = 0.9999999999;//copy from var with 1.0
cout << setprecision(2);
cout << fixed;
cout << "precision=2 fix :" << 1.1111111111111 << ", json: " << testj<< endl;

What is the expected behavior?

it should print
precision=2 fix :1.11, json: {"data1":7.00, etc... ,"data2":1.00}

And what is the actual behavior instead?

but the output is
precision=2 fix :1.11, json: {"data1":7.0000000001, etc... ,"data2":0.9999999999}
i try to test setw and setfill, it works; but doesn't work for setprecision

Which compiler and operating system are you using?

• Compiler: gcc.7.3.0
• Operating system: suse11.3

Which version of the library did you use?

• latest release version 3.9.1
• other release - please state the version: 3.7.3
• the develop branch

[nlohmann]

It's not really a bug - the library has no option to change the output of floating-point numbers. If you use testj.get<double>() instead of testj, then it should work.

[young66]

sorry，i show this code ,just want to say setprecision doesn't work.
my code is like this, i have update the code in issue.

json testj;
testj["data1"] = 7.0000000001;//copy from var like 7.0
//some more data in json
testj["data2"] = 0.9999999999;//copy from var like 1.0
cout << setprecision(2);
cout << fixed;
cout << "precision=2 fix :" << 1.1111111111111 << ", json: " << testj<< endl;

it should print
precision=2 fix :1.11, json: {"data1":7.00, etc... ,"data2":1.00}

but the output is
precision=2 fix :1.11, json: {"data1":7.0000000001, etc... ,"data2":0.9999999999}

i think o.precision() should pass to the function operator<< in include/nlohmann/json.hpp
and use the precision to format the double/float var for function dump_float in single_include/nlohmann/json.hpp
like this

friend std::ostream& operator<<(std::ostream& o, const basic_json& j)
{
// read width member and use it as indentation parameter if nonzero
const bool pretty_print = o.width() > 0;
const auto indentation = pretty_print ? o.width() : 0;
// reset width to 0 for subsequent calls to this stream
o.width(0);

// do the actual serialization // <<-------o.fill() has passed to serializer, so do o.precision for double formating
serializer s(detail::output_adapter<char>(o), o.fill(), o.precision());
s.dump(j, pretty_print, false, static_cast<unsigned int>(indentation));
return o;
}

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[axtschmied]

Hello!

This topic has been closed due to inactivity, but it would be a very useful feature in my opinion as well.

[DinoMalpera]

We also need this. It would be very useful.