views:

364

answers:

1

I am writing a C++ app that has to parse a date/time string and give out epoch time. But the format of the date/time string can be more than one (actually say 25 formats) like

    "EEE, dd MMM yyyy HH:mm:ss '('ZZZ')'",
    "EEE, dd MMM yyyy HH:mm:ss '\"'ZZZ'\"'",
    "EEE, dd MMM yyyy hh:mm:ss z",
    "EEE, dd MMM yyyy HH:mm Z",
    "EEE, dd MMM yyyy HH:mm:ss",
    "EEE, MMM dd, yyyy 'at' HH:mm:ss z",
    "EEE M/dd/yyyy hh:mm:ss a",
    "EEE MMM dd HH:mm:ss z yyyy",
    "EEE MMM dd hh:mm:ss yyyy",
    "EEEE, MMMM dd, yyyy hh:mm:ss a",
    "EEEE, MMMM dd, yyyy HH:mm a",
    "EEEE, MMMM dd, yyyy HH:mm",
    "MMM dd, yyyy hh:mm:ss a",
    "dd MMM yyyy hh:mm:ss z",
    "dd-MMM-yyyy HH:mm:ss z",
    "dd MMM yy HH:mm:ss",
    "MM/dd/yyyy  hh:mm a  (EEEE)",
    "MM/dd/yyyy hh:mm a (EEEE)",
    "MM/dd/yyyy hh:mm:ss",
    "MM/dd/yyyy hh:mm a Z",
    "MM/dd/yyyy hh:mma Z",
    "MM/dd/yyyy hh:mma",
    "MM/dd/yyyy hh:mm a",
    "MM/dd/yyyy hh:mm Z",
    "MM/dd/yy hh:mm a Z",
    "MM/dd/yy hh:mma Z",
    "MM/dd/yy HH:mm a",
    "MM/dd/yy HH:mm Z",
    "MM/dd/yyyy",   
    "yyyy-MM-dd HH:mm:ss",
    "yyyyMMddhhmmss",
    "yyyyMMddhhmm",
    "yyyyMMdd"

Now, i need to take the string, figure out it belongs to which of these formats, then get the time in epoch.

Can you suggest a way to do this. Code samples will be really helpful. I am banking on Boost libraries. Let me know if this can be achieved by boost date/time parsing libraries.

Thanks in advance, AJ

+3  A: 

I suppose you could attempt to convert your string into ptime using each of these formats and pick the ones that do not result in a not_a_date_time.:

The boost format flags are slightly different from yours, I'll do just the last five for this example:

#include <iostream>
#include <boost/date_time/posix_time/posix_time.hpp>
using boost::posix_time::time_input_facet;
using std::locale;
const locale inputs[] = {
    locale(locale::classic(), new time_input_facet("%m/%d/%Y")),
    locale(locale::classic(), new time_input_facet("%Y-%m-%d %H:%M:%S")),
    locale(locale::classic(), new time_input_facet("%Y%m%d%H%M%S")),
    locale(locale::classic(), new time_input_facet("%Y%m%d%H%M")),
    locale(locale::classic(), new time_input_facet("%Y%m%d")) };
const size_t formats = sizeof(inputs)/sizeof(inputs[0]);

time_t ptime_to_time_t(boost::posix_time::ptime t)
{
       static boost::posix_time::ptime epoch(boost::gregorian::date(1970,1,1));
       return (t-epoch).ticks() / boost::posix_time::time_duration::ticks_per_second();
}
int main()
{
       std::string msg = "2010-08-04 08:34:12";

       for(size_t i=0; i<formats; ++i)
       {
           std::istringstream ss(msg);
           ss.imbue(inputs[i]);
           boost::posix_time::ptime this_time;
           ss >> this_time;

           if(this_time != boost::posix_time::not_a_date_time)
               std::cout << this_time << " or " << ptime_to_time_t(this_time) << std::endl;
       }
}
Cubbi
Thanks cubbi for taking so much time, i will try this and get back.
AJ
Now, it looks like %d and %m have to have 2 characters, ie.. it has to be like 01 instead of 1 and 05 instead of 5. How can i handle that case. This seems to be causing an issue for me.Looked at the format flags could not figure out much.
AJ
If your days and months can be single digits, with no preceding zero or space, then some of these formats cannot be parsed unambiguously. For `yyyyMMdd`, would 2010111 be 2010-01-11 or 2010-11-01? In any case, if there's something date_time IO cannot handle, there's always boost.spirit.
Cubbi
Any idea how i take into account the timezones as well during input?!
AJ