[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: NcML Aggregation



Joe McLean wrote:

Hi John,

I have an aggregation server (based on NcML) working at
http://www.ferret.noaa.gov:9090/dodsC.

The variables in the source URLS are shaped on the existing dimension
'recnum'.  I have added two more dimensions/coordinate variables to the
Netcdf Dataset. They are  'lat' and 'lon', each are length="1".  Do you
know of a way to use the NcML Dataset specification to add these two
dimensions to the shape of an existing variable in the file?  Although
documentation explains readMetadata and rename elements, there is no
reshape element.  I'm guessing that is on purpose (reshaping variables
asks for lots of problems), but maybe there is a work-around.

Joe

PS: I've been bugging you so much, is there a mailing list I should be
on?
---------------------------

Hi Joe.

I think that you dont want to add lat, lon to the existing variables, since that would make them into a grid. If I understand your data correctly, what you have is station data. So, if the lat, lon doesnt change for the station, then lat(1), lon(1) is right. If it changes, then the right thing to do is to add lat(recnum), lon(recnum) variables, which list the lat and lon for each measurement.

Actually, this depends on the conventions for describing station data, so you might want to run that by Steve Hankin and others who have beeen working on station data. We are starting to work on conventions in THREDDS and netCDF for station data, so I'd like to continue to be part of the conversation.