Hi Bob,
I totally agree that our files must be CF-compliant. But we want to be
able to read the files with most programs, and the whole CF can be
difficult to implement. The change I made is still CF compliant and it
works with the netcdf-java, too. Yes, netcdf-java only implements a
subset of CF, but since I'm not funding them, I can live with that subset.
Writing a special format is usually simple (and we have some flexibility
within CF), reading all formats is hard, so we're trying to find a
compromise.
BTW since you have a NOAA email address: we started from the NOAA
templates:
https://geo-ide.noaa.gov/wiki/index.php?title=NODC_NetCDF_Trajectory_Template
(and eliminated some conflicts with CF from them, e.g. axis)
Best regards,
Heiko
On 2015-08-27 21:21, Bob Simons - NOAA Federal wrote:
> I'm not following this closely, but I was bothered by "we now know how
> to create a netcdf-java compliant trajectory file". netcdf-java is just
> one tool. There are other tools now, and there will be more in the
> future. netcdf-java probably supports a subset/superset of the file
> types defined by the CF DSG standard (which allows a huge number of
> minor variations).
>
> I think a far better goal is to seek to make files which are compliant
> with the CF DSG standard, and thus with any tool that follows the CF DSG
> standard. If a given tool (e.g., netcdf-java) can't deal with a valid CF
> DSG file, then it is a bug in that tool and you can
> legitimately/reasonably ask the developer to fix the bug.
>
> On Thu, Aug 27, 2015 at 12:01 PM, Heiko Klein <Heiko.Klein@xxxxxx
> <mailto:Heiko.Klein@xxxxxx>> wrote:
>
> Hi Kyle,
>
> I made now the time-dimension 2-dimensional, i.e. time(obs, trajectory),
> removing the one simplification made below. netcdf-java seems now to be
> able to read the trajectories. Since the page:
>
> http://www.unidata.ucar.edu/software/thredds/current/netcdf-java/reference/FeatureDatasets/CFpointImplement.html
> already mentions differences from CF, maybe this one could be entered
> there, too? (Or maybe it's easy to extend the cdm to support
> one-dimensional time-axes?)
>
>
>
> I wasn't able to plot trajectories as seen in netcdf-java 2.2. Is it
> just me not finding the right buttons or was this feature removed? This
> is not critical, but would be nice for debugging.
>
>
> But the most important point is that we now know how to create a
> netcdf-java compliant trajectory file. Thanks for your help.
>
> Heiko
>
>
> On 2015-08-27 20:17, Heiko Klein wrote:
> > Hi Kyle,
> >
> > we're trying to use quite strictly H.4.1 here, with one addition:
> >
> > Since all trajectories have the same timesteps we use 'time'='obs' and
> > we don't have a special obs dimension. According to CF, this is
> allowed:
> > (Chapter H.4.1: "If the trajectories all have the same set of
> times, the
> > time auxiliary coordinate variable could be one-dimensional time(obs),
> > or replaced by a one-dimensional coordinate variable time(time),
> where
> > the size of the time dimension is now equal to the number of
> elements of
> > each trajectory. In the latter case, listing the time coordinate
> > variable in the coordinates attribute is optional.")
> >
> > Best regards,
> >
> > Heiko
> >
> >
> > On 2015-08-27 16:22, Kyle Wilcox wrote:
> >> Hi Heiko,
> >>
> >> At first glace, I can't figure out which CF specification you were
> >> trying to adhere to: H.4.1, H.4.2, H.4.3, or H.4.4. It appears
> you may
> >> have tried to adopt H.4.2 to use multiple trajectories, which I'm not
> >> sure is supported in NetCDF-Java.
> >>
> >> Take a look at these Trajectory
> >> examples:
>
> https://github.com/asascience-open/CFPointConventions/tree/master/trajectory.
> >> Each folder contains a .cdl, .nc (that will load into
> NetCDF-Java), and
> >> an .ncml file.
> >>
> >> Hope that helps,
> >> Kyle
> >>
> >>
> >> On Thu, Aug 27, 2015 at 4:21 AM, Heiko Klein <Heiko.Klein@xxxxxx
> <mailto:Heiko.Klein@xxxxxx>
> >> <mailto:Heiko.Klein@xxxxxx <mailto:Heiko.Klein@xxxxxx>>> wrote:
> >>
> >> Hi,
> >>
> >> I'm trying to read trajectory data with netcdf-java. I've
> seen figures
> >> of this with netcdf-java 2.2 at
> >>
> >>
>
> https://www.unidata.ucar.edu/software/thredds/v4.3/netcdf-java/tutorial/PointDatatype.html
> >>
> >> but both netcdf-4.6 and CF-1.6 have a different layout for
> >> trajectory data.
> >>
> >>
> >>
> >> I've tried now to generate a netcdf/CF file following the
> guidelines at
> >>
>
> http://www.unidata.ucar.edu/software/thredds/current/netcdf-java/reference/FeatureDatasets/CFpointImplement.html
> >> (see attached file).
> >>
> >>
> >> ToolsUI gives me a 'CoordSys' with trajectory, time, lat, lon and
> >> time,lat,lon have the expected axisTypes. Using FeatureScan,
> I get a
> >> coverage of 'Swath(T,Y,X): trajectory', which looks ok, but I
> get also
> >> an error message:
> >>
> >> ---------
> >> /home/heikok/openoil2.nc4
> >> NetCDF-4
> >> map = 'f:D(2)->R(4)'
> >>
> >> null
> >>
> >> CoverageFactory for '/home/heikok/openoil2.nc4'
> >> coverage = Swath
> >> FeatureDatasetFactoryManager errlog = Table
> >> PsuedoStructure(trajectory)/MultidimPseudo(trajectory,time)
> featureType
> >> TRAJECTORY: lat/lon/time coord not found
> >> **Failed to find FeatureDatasetFactory for=
> /home/heikok/openoil2.nc4
> >> datatype=TRAJECTORY
> >> ----------
> >>
> >>
> >> According to 'CoordSys', lat/lon/time are availble.
> >>
> >> The PointFeature/TRAJECTORY wasn't then able to plot the
> data. Any idea
> >> how I can correct my data to enable plotting in netcdf-java UI?
> >>
> >>
> >> Best regards,
> >>
> >> Heiko
> >>
> >>
> >>
> >> _______________________________________________
> >> netcdf-java mailing list
> >> netcdf-java@xxxxxxxxxxxxxxxx
> <mailto:netcdf-java@xxxxxxxxxxxxxxxx>
> <mailto:netcdf-java@xxxxxxxxxxxxxxxx
> <mailto:netcdf-java@xxxxxxxxxxxxxxxx>>
> >> For list information or to unsubscribe, visit:
> >> http://www.unidata.ucar.edu/mailing_lists/
> >>
> >>
> >
>
> _______________________________________________
> netcdf-java mailing list
> netcdf-java@xxxxxxxxxxxxxxxx <mailto:netcdf-java@xxxxxxxxxxxxxxxx>
> For list information or to unsubscribe, visit:
> http://www.unidata.ucar.edu/mailing_lists/
>
>
>
>
> --
> Sincerely,
>
> Bob Simons
> IT Specialist
> Environmental Research Division
> NOAA Southwest Fisheries Science Center
> 99 Pacific St., Suite 255A (New!)
> Monterey, CA 93940 (New!)
> Phone: (831)333-9878 (New!)
> Fax: (831)648-8440
> Email: bob.simons@xxxxxxxx <mailto:bob.simons@xxxxxxxx>
>
> The contents of this message are mine personally and
> do not necessarily reflect any position of the
> Government or the National Oceanic and Atmospheric Administration.
> <>< <>< <>< <>< <>< <>< <>< <>< <><
>
--
Dr. Heiko Klein Norwegian Meteorological Institute
Tel. + 47 22 96 32 58 P.O. Box 43 Blindern
http://www.met.no 0313 Oslo NORWAY
