This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
> On Thu, Apr 30, 2009 at 1:20 PM, Unidata THREDDS Support > <address@hidden> wrote: > >> NetCDF-Java Gurus, > >> > >> I may have hit the first ocean model that I can't use NcML to CF-ize. > >> > >> Output from the Harvard Ocean Prediction System (example URL: > >> http://coast-enviro.er.usgs.gov/thredds/dodsC/models/share/hops.nc.dds > >> > >> stores both lon and lat in a single variable, and u,v in a single > >> variable, by adding an extra dimension. > >> So u and v is stored in: > >> > >> Float32 vtot[time = 17][vlat = 138][vlon = 107][outlev = 30][vector = 2]; > >> > >> Can we use NcML to split u [vector = 1] and v [vector = 2] into > >> separate variables? > >> > >> If not, would this be (a) easy (2 hours); (b); medium (2 days); or (c) > >> difficult (2 months)? > > > > the best thing is to write a CoordSysBuilder to fix the problem. AWIPS > > files have same problem, thats how we fix them. > > > > I will eventually add a feature for NcML to split data like this, but its > > not possible now. > > > > Its a few hours work by a java programmer. is there someone who can do it > > there? > > Which is a few hours work? Adding the feature for NcML or writing a > new CoordSysBuilder (I'm pretty sure it's the latter, but just want to > make sure!) sorry, the latter Ticket Details =================== Ticket ID: FIZ-452444 Department: Support THREDDS Priority: Normal Status: Closed