Hi Daryl,
It looks like the name change were due to an update to the NCEP grib
tables. We use the grib tables that the centers provide, and it looks
like their update included some trivial (capital letters) and non-trival
(accumulation time definition) changes. netCDF-Java (and thus TDS) tries
to be a "center agnostic" grib reader, and to do that we rely on the
tables supplied from the centers generating the products, so when those
change, our stuff changes.
The URLs for the netCDF Subset Service did change slightly, and this
was included in the announcement on the thredds@xxxxxxxxxxxxxxxx list.
See
http://www.unidata.ucar.edu/software/thredds/v4.5/tds/UpgradingTo4.5.html for
more details.
Cheers,
Sean
On 3/31/14, 12:03 PM, daryl herzmann wrote:
On Mon, 31 Mar 2014, Sean Arms wrote:
Which datasets did you observe variable name changes? I'll look into
it to see why they changed. Again, as far as I know the only reason a
variable name would have changed is due to an updated grib table
obtained from a data provider.
Here was a diff I did to my ingestor on Friday
https://github.com/akrherz/iem/commit/ec2f1bb0cd193635a897bb088f7954db74bb3694
- Convective_precipitation_surface_3_Hour_Accumulation
+ Convective_precipitation_surface_Mixed_Intervals_Accumulation
-
Large-scale_precipitation_non-convective_surface_Mixed_intervals_Accumulation
+
Large-scale_precipitation_non-convective_surface_Mixed_Intervals_Accumulation
(I is now capitolized)
The change also noted slight differences in the file paths... (see
first part of that shown diff)
daryl
