I am trying to write a simple timestamping system that provides epoch seconds and fractional seconds from the current time. I am using boost library and have something like this:
const boost::posix_time::ptime epoch(boost::gregorian::date(1970, 1, 1));
boost::posix_time::ptime time() {
boost::posix_time::ptime now = boost::posix_time::microsec_clock::universal_time();
return now;
}
boost::posix_time::time_duration dur = (time() - epoch);
and then use the following elements to extract the epoch values:
dur.total_seconds();
dur.fractional_seconds();
Specifically, will this return a proper unix time? If not, any suggestions as to how to correct it? Thanks.
Yes, that should work, but, to be certain, there's always experimental evidence:
#include <iostream>
#include <time.h>
#include <boost/date_time.hpp>
namespace bpt = boost::posix_time;
namespace bg = boost::gregorian;
int main()
{
bpt::time_duration dur = bpt::microsec_clock::universal_time()
- bpt::ptime(bg::date(1970, 1, 1));
timespec ts;
clock_gettime(CLOCK_REALTIME, &ts);
std::cout << std::setfill('0')
<< " boost: " << dur.total_seconds() << '.' << std::setw(6)
<< dur.fractional_seconds() << '\n'
<< " ctime: " << time(NULL) << '\n'
<< " posix: " << ts.tv_sec << '.' << std::setw(9)
<< ts.tv_nsec << '\n';
}
I get
Linux/gcc
boost: 1361502964.664746
ctime: 1361502964
posix: 1361502964.664818326
Sun/Sun Studio
boost: 1361503762.775609
ctime: 1361503762
posix: 1361503762.775661600
AIX/XLC
boost: 1361503891.342930
ctime: 1361503891
posix: 1361503891.342946000
and even Windows/Visual Studio
boost: 1361504377.084231
ctime: 1361504377
Looks like they all agree on how many seconds passed since date(1970,1,1)
Related
How do I get a uint unix timestamp in C++? I've googled a bit and it seems that most methods are looking for more convoluted ways to represent time. Can't I just get it as a uint?
C++20 introduced a guarantee that time_since_epoch is relative to the UNIX epoch, and cppreference.com gives an example that I've distilled to the relevant code, and changed to units of seconds rather than hours:
#include <iostream>
#include <chrono>
int main()
{
const auto p1 = std::chrono::system_clock::now();
std::cout << "seconds since epoch: "
<< std::chrono::duration_cast<std::chrono::seconds>(
p1.time_since_epoch()).count() << '\n';
}
Using C++17 or earlier, time() is the simplest function - seconds since Epoch, which for Linux and UNIX at least would be the UNIX epoch. Linux manpage here.
The cppreference page linked above gives this example:
#include <ctime>
#include <iostream>
int main()
{
std::time_t result = std::time(nullptr);
std::cout << std::asctime(std::localtime(&result))
<< result << " seconds since the Epoch\n";
}
#include<iostream>
#include<ctime>
int main()
{
std::time_t t = std::time(0); // t is an integer type
std::cout << t << " seconds since 01-Jan-1970\n";
return 0;
}
The most common advice is wrong, you can't just rely on time(). That's used for relative timing: ISO C++ doesn't specify that 1970-01-01T00:00Z is time_t(0)
What's worse is that you can't easily figure it out, either. Sure, you can find the calendar date of time_t(0) with gmtime, but what are you going to do if that's 2000-01-01T00:00Z ? How many seconds were there between 1970-01-01T00:00Z and 2000-01-01T00:00Z? It's certainly no multiple of 60, due to leap seconds.
As this is the first result on google and there's no C++20 answer yet, here's how to use std::chrono to do this:
#include <chrono>
//...
using namespace std::chrono;
int64_t timestamp = duration_cast<milliseconds>(system_clock::now().time_since_epoch()).count();
In versions of C++ before 20, system_clock's epoch being Unix epoch is a de-facto convention, but it's not standardized. If you're not on C++20, use at your own risk.
#include <iostream>
#include <sys/time.h>
using namespace std;
int main ()
{
unsigned long int sec= time(NULL);
cout<<sec<<endl;
}
I created a global define with more information:
#include <iostream>
#include <ctime>
#include <iomanip>
#define __FILENAME__ (__builtin_strrchr(__FILE__, '/') ? __builtin_strrchr(__FILE__, '/') + 1 : __FILE__) // only show filename and not it's path (less clutter)
#define INFO std::cout << std::put_time(std::localtime(&time_now), "%y-%m-%d %OH:%OM:%OS") << " [INFO] " << __FILENAME__ << "(" << __FUNCTION__ << ":" << __LINE__ << ") >> "
#define ERROR std::cout << std::put_time(std::localtime(&time_now), "%y-%m-%d %OH:%OM:%OS") << " [ERROR] " << __FILENAME__ << "(" << __FUNCTION__ << ":" << __LINE__ << ") >> "
static std::time_t time_now = std::time(nullptr);
Use it like this:
INFO << "Hello world" << std::endl;
ERROR << "Goodbye world" << std::endl;
Sample output:
16-06-23 21:33:19 [INFO] main.cpp(main:6) >> Hello world
16-06-23 21:33:19 [ERROR] main.cpp(main:7) >> Goodbye world
Put these lines in your header file. I find this very useful for debugging, etc.
Windows uses a different epoch and time units: see
Convert Windows Filetime to second in Unix/Linux
What std::time() returns on Windows is (as yet) unknown to me (;-))
While I realize this is probably one of many identical questions, I can't seem to figure out how to properly use std::chrono. This is the solution I cobbled together.
#include <stdlib.h>
#include <iostream>
#include <chrono>
typedef std::chrono::high_resolution_clock Time;
typedef std::chrono::milliseconds ms;
float startTime;
float getCurrentTime();
int main () {
startTime = getCurrentTime();
std::cout << "Start Time: " << startTime << "\n";
while(true) {
std::cout << getCurrentTime() - startTime << "\n";
}
return EXIT_SUCCESS;
}
float getCurrentTime() {
auto now = Time::now();
return std::chrono::duration_cast<ms>(now.time_since_epoch()).count() / 1000;
}
For some reason, this only ever returns integer values as the difference, which increments upwards at rate of 1 per second, but starting from an arbitrary, often negative, value.
What am I doing wrong? Is there a better way of doing this?
Don't escape the chrono type system until you absolutely have to. That means don't use .count() except for I/O or interacting with legacy API.
This translates to: Don't use float as time_point.
Don't bother with high_resolution_clock. This is always a typedef to either system_clock or steady_clock. For more portable code, choose one of the latter.
.
#include <iostream>
#include <chrono>
using Time = std::chrono::steady_clock;
using ms = std::chrono::milliseconds;
To start, you're going to need a duration with a representation of float and the units of seconds. This is how you do that:
using float_sec = std::chrono::duration<float>;
Next you need a time_point which uses Time as the clock, and float_sec as its duration:
using float_time_point = std::chrono::time_point<Time, float_sec>;
Now your getCurrentTime() can just return Time::now(). No fuss, no muss:
float_time_point
getCurrentTime() {
return Time::now();
}
Your main, because it has to do the I/O, is responsible for unpacking the chrono types into scalars so that it can print them:
int main () {
auto startTime = getCurrentTime();
std::cout << "Start Time: " << startTime.time_since_epoch().count() << "\n";
while(true) {
std::cout << (getCurrentTime() - startTime).count() << "\n";
}
}
This program does a similar thing. Hopefully it shows some of the capabilities (and methodology) of std::chrono:
#include <iostream>
#include <chrono>
#include <thread>
int main()
{
using namespace std::literals;
namespace chrono = std::chrono;
using clock_type = chrono::high_resolution_clock;
auto start = clock_type::now();
for(;;) {
auto first = clock_type::now();
// note use of literal - this is c++14
std::this_thread::sleep_for(500ms);
// c++11 would be this:
// std::this_thread::sleep_for(chrono::milliseconds(500));
auto last = clock_type::now();
auto interval = last - first;
auto total = last - start;
// integer cast
std::cout << "we just slept for " << chrono::duration_cast<chrono::milliseconds>(interval).count() << "ms\n";
// another integer cast
std::cout << "also known as " << chrono::duration_cast<chrono::nanoseconds>(interval).count() << "ns\n";
// floating point cast
using seconds_fp = chrono::duration<double, chrono::seconds::period>;
std::cout << "which is " << chrono::duration_cast<seconds_fp>(interval).count() << " seconds\n";
std::cout << " total time wasted: " << chrono::duration_cast<chrono::milliseconds>(total).count() << "ms\n";
std::cout << " in seconds: " << chrono::duration_cast<seconds_fp>(total).count() << "s\n";
std::cout << std::endl;
}
return 0;
}
example output:
we just slept for 503ms
also known as 503144616ns
which is 0.503145 seconds
total time wasted: 503ms
in seconds: 0.503145s
we just slept for 500ms
also known as 500799185ns
which is 0.500799 seconds
total time wasted: 1004ms
in seconds: 1.00405s
we just slept for 505ms
also known as 505114589ns
which is 0.505115 seconds
total time wasted: 1509ms
in seconds: 1.50923s
we just slept for 502ms
also known as 502478275ns
which is 0.502478 seconds
total time wasted: 2011ms
in seconds: 2.01183s
Is there a simpler way to do a time stamp for saving a file/creating a directory as a date time stamp ?
only using standard library (not boost). Is there a faster way to do it ?
This is my current code
std::chrono::system_clock::time_point now = std::chrono::system_clock::now();
std::time_t tt = std::chrono::system_clock::to_time_t(now);
tm utc_tm = *gmtime(&tt);
oname.str("");
oname << (utc_tm.tm_year + 1900) << '-' << std::setfill('0') << std::setw(2) << (utc_tm.tm_mon + 1) << '-' << std::setfill('0') << std::setw(2) << utc_tm.tm_mday << " " << std::setfill('0') << std::setw(2)<< utc_tm.tm_hour <<':' << std::setfill('0') << std::setw(2) << utc_tm.tm_min <<':' << std::setfill('0') << std::setw(2) << utc_tm.tm_sec;
ts = oname.str();
There is a less tortuous way:
#include <string>
#include <ctime>
std::string get_timestamp()
{
auto now = std::time(nullptr);
char buf[sizeof("YYYY-MM-DD HH:MM:SS")];
return std::string(buf,buf +
std::strftime(buf,sizeof(buf),"%F %T",std::gmtime(&now)));
}
It is very probably also faster, because it is less tortuous, but that is
also very probably immaterial in a setting where disc I/O is in play.
This gives you the same timestamps as your own code, e.g.
2015-03-28 10:48:45
See std::time and
std::strftime to
understand how the desired formatting is achieved and note that std::strftime
returns the length of the string it has composed, excluding its nul-terminator.
This code is standard, but if you are working with MS VC++ 2013 or later then
you could also consider the use of std::put_time,
as in:
#include <iomanip>
#include <sstream>
#include <string>
#include <ctime>
std::string get_timestamp()
{
auto now = std::time(nullptr);
std::ostringstream os;
os << std::put_time(std::gmtime(&now),"%F %T");
return os.str();
}
which is simpler still. (I have not tested that.) std::put_time however
is unsupported by gcc as of 4.9.
Seemingly you want your timestamps formatted as YYYY-MM-DD HH:MM:SS. If they
are to be used in filenames, it would be more prudent to keep them free of spaces:
perhaps YYYY-MM-DD_HH:MM:SS.
When I run this in my ubuntu, I get the bottom result. What could go wrong with this? There is no way the result year is 1970, and see when I do the reconvert the ms back, the result aren't the same. But when I run it at http://coliru.stacked-crooked.com/ it was fine.
#include <boost/date_time/gregorian/gregorian.hpp>
#include <boost/date_time/posix_time/posix_time.hpp>
#include <string>
#include <iostream>
int main()
{
using namespace boost::posix_time;
ptime time_t_epoch(boost::gregorian::date(1970,1,1));
auto ms = (boost::posix_time::microsec_clock::local_time() - time_t_epoch).total_microseconds();
std::cout << "ms: " << ms << std::endl;
std::cout << "ms: " << boost::posix_time::microseconds(ms).total_microseconds() << std::endl;
ptime newTime = time_t_epoch + boost::posix_time::microseconds(ms);
std::cout << boost::posix_time::to_iso_string(time_t_epoch) << std::endl;
std::cout << boost::posix_time::to_iso_string(newTime) << std::endl;
std::stringstream ss;
ss << newTime;
std::cout << ss.str()<< std::endl;
}
Result:
ms: 1424976249761218
ms: 4576956085582
19700101T000000
19700222T232236.085582
1970-Feb-22 23:22:36.085582
Machine environment = Ubuntu 12 + cmake 2.8.7 + boost 1.48.0.2
1970/01/01 + 4576956085582 microseconds -> 1970-Feb-22 is correct
1970/01/01 + 1424976249761218 microseconds -> 2015-Feb-26 so ms is presumably correct
What is not correct is that
boost::posix_time::microseconds(ms).total_microseconds() != ms
It might be related to this bug https://svn.boost.org/trac/boost/ticket/3487 though I'm not sure.
The following piece of code is used to print the time in the logs:
#define PRINTTIME() struct tm * tmptime;
time_t tmpGetTime;
time(&tmpGetTime);
tmptime = localtime(&tmpGetTime);
cout << tmptime->tm_mday << "/" <<tmptime->tm_mon+1 << "/" << 1900+tmptime->tm_year << " " << tmptime->tm_hour << ":" << tmptime->tm_min << ":" << tmptime->tm_sec<<">>";
Is there any way to add milliseconds to this?
To have millisecond precision you have to use system calls specific to your OS.
In Linux you can use
#include <sys/time.h>
timeval tv;
gettimeofday(&tv, 0);
// then convert struct tv to your needed ms precision
timeval has microsecond precision.
In Windows you can use:
#include <Windows.h>
SYSTEMTIME st;
GetSystemTime(&st);
// then convert st to your precision needs
Of course you can use Boost to do that for you :)
//C++11 Style:
cout << "Time in Milliseconds =" <<
chrono::duration_cast<chrono::milliseconds>(chrono::steady_clock::now().time_since_epoch()).count()
<< std::endl;
cout << "Time in MicroSeconds=" <<
chrono::duration_cast<chrono::microseconds>(chrono::steady_clock::now().time_since_epoch()).count()
<< std::endl;
You need a timer with a higher resolution in order to capture milliseconds. Try this:
int cloc = clock();
//do something that takes a few milliseconds
cout << (clock() - cloc) << endl;
This is of course dependent on your OS.
The high resolution timers are usually gettimeofday on Linux style platforms and QueryPerformanceCounter on Windows.
You should be aware that timing the duration of a single operation (even with a high resolution timer) will not yield accurate results. There are too many random factors at play. To get reliable timing information, you should run the task to be timed in a loop and compute the average task time. For this type of timing, the clock() function should be sufficient.
If you don't want to use any OS-specific code, you can use the ACE package which supplies the ACE_OS::gettimeofday function for most standard operating systems.
For example:
ACE_Time_Value startTime = ACE_OS::gettimeofday();
do_something();
ACE_Time_Value endTime = ACE_OS::gettimeofday();
cout << "Elapsed time: " << (endTime.sec() - startTime.sec()) << " seconds and " << double(endTime.usec() - startTime.usec()) / 1000 << " milliseconds." << endl;
This code will work regardless of your OS (as long as ACE supports this OS).
In Ubuntu 16.04 this worked for me...
const std::string currentDateTime() {
char fmt[64], buf[64];
struct timeval tv;
struct tm *tm;
gettimeofday(&tv, NULL);
tm = localtime(&tv.tv_sec);
strftime(fmt, sizeof fmt, "%Y-%m-%d %H:%M:%S.%%06u", tm);
snprintf(buf, sizeof buf, fmt, tv.tv_usec);
return buf;
}
Then, with...
std::cout << currentDateTime();
I get...
2016-12-29 11:09:55.331008
New answer for old question using C++11 or C++14 and this free, open-source library:
#include "tz.h"
#include <iostream>
int
main()
{
using namespace date;
using namespace std;
using namespace std::chrono;
auto now = make_zoned(current_zone(), floor<milliseconds>(system_clock::now()));
cout << format("%e/%m/%Y %T", now) << '\n';
}
This just output for me:
16/01/2017 15:34:32.167
which is my current local date and time to millisecond precision. By eliminating the floor<milliseconds>() you will automatically get whatever precision your system_clock has.
If you wanted the result as a UTC timestamp instead of a local timestamp, it is even easier:
auto now = floor<milliseconds>(system_clock::now());
cout << format("%e/%m/%Y %T", now) << '\n';
And if you want a UTC timestamp and you aren't picky about the precision or the format, you can just:
cout << system_clock::now() << '\n';
which just output for me:
2017-01-16 20:42:11.267245