Re: FW: [netCDFJava #RUG-895669]: NetCDF Java API

Hi Ross:

Sorry for the long delay.

GRIB has a notion of "reference time" and "valid time". We use the 
earliest reference time in the collection for the time units string. The 
problem is that "reference time" doesnt always mean "run time", but if 
in your particular case it does, then you can use it. heres a test routine:

  public static void main(String[] args) {
    String s = "Hour since 2007-03-22T00:00:00Z";
    CalendarDateUnit  cdu = CalendarDateUnit.of(null, s);
    System.out.printf("%s == %s%n base = %s or %s%n", s, cdu, 
cdu.getBaseCalendarDate(), cdu.getBaseDate());
  }

One of the reasons it took me so long to answer is that Ive been 
considering adding an auxiliary time coordinate that would document the 
reference time. However, at this point its not going to happen soon.

thanks for your patience.

John

On 12/17/2012 2:45 AM, Callister, Ross wrote:
> Hi John,
>
> Season's Greetings!
>
> Please could you let me know if you've had chance to look at this one yet?
>
> All I really need to know now is if there is a way of extracting the Run Date 
> from the Grib file using the api?
>
> The only way I can see of doing it currently is to parse the String returned by the units 
> attribute of the time variable.  This is working for me, but is a bit ugly and 
> inefficient (i.e. converting "Hour since 2012-11-07T00:00:00.000Z" to a 
> Calendar object).
>
> Thanks,
>
> Ross Callister
> Met Office   FitzRoy Road   Exeter   Devon   EX1 3PB  United Kingdom
> Tel: +44 (0)1392 886578  Mobile: +44 (0)7841 098979
> Email: address@hidden, address@hidden
>
>
>
> -----Original Message-----
> From: Unidata netCDF Java Support [mailto:address@hidden]
> Sent: 26 November 2012 21:36
> To: Callister, Ross
> Cc: Rowe, Paul
> Subject: [netCDFJava #RUG-895669]: NetCDF Java API
>
> Hi Ross,
>
> Sorry for the delay. We've been pretty focused the last few weeks on making 4.3 
> stable release and had our Thanksgiving holiday last week. Unfortunately, we 
> have a two day planning meeting tomorrow and Wednesday. So probably won't get 
> back to you on this issue until later this week.
>
> I'm assigning this ticket to John as he has done all our recent GRIB work.
>
> Ethan
>
> PS Say Hi to Pete for me.
>
> Ross Callister wrote:
> > Hi guys,
> >
> > I am currently working at the UK Met Office looking at reading GRIB1
> > files using your Java api (4.3.8 Beta).
> >
> > It seems to work well, although I have run into a problem when trying
> > to access the Model Run Time for variables in the GRIB file.  I have
> > seen this data in the output of the NetcdfFile.getDetailedInfo()
> > method and it seems to come from the Grib1Iosp>  GribCollection>
> > GroupHcs>  TimeCoord class route (the runDate variable of the
> > TimeCoord class).  I need to know if it is possible to get the runDate
> > associated with a given variable using the API?
> >
> > Also, if you have a file that has multiple GRIB run dates and or
> > co-ordinates, how do you differentiate between the variables e.g. what
> > would the following code return:
> >
> > Variable timeVar = findVariable(dataFile, "time");
> >
> > i.e. which variable's time would it fetch?
> >
> >
> > Thanks in advance,
> >
> > Ross Callister
> >
> >
> > p.s. I'm working closely with Pete Trevelyan who I believe you know.
>
> Ticket Details
> ===================
> Ticket ID: RUG-895669
> Department: Support netCDF Java
> Priority: Urgent
> Status: Open