I see now that simply removing the first part of the URL
https://oceanwatch.pfeg.noaa.gov/thredds/remoteCatalogService?catalog=
isn't a great solution.
Previously, the datasets in the remote catalogs appeared to be in the local
thredds
e.g., https://oceanwatch.pfeg.noaa.gov/thredds/dodsC/satellite/MPOC/1day
but removing the first part of the URL causes the datasets to have URLs in
the remote thredds
e.g., https://thredds1.pfeg.noaa.gov/thredds/dodsC/satellite/MPOC/1day
That isn't good because the remote thredds may be a semi-private thredds
that the administrator would prefer not to have addressed directly (e.g.,
because the url may change periodically, e.g., for load balancing).
So it would be nice if netcdf-java worked as it used to, where catalogs
that point to remote catalogs still return datasets that appear to be in
the local THREDDS.
Thank you for investigating.
https://thredds1.pfeg.noaa.gov/thredds/dodsC/satellite/MPOC/1day[end]
On Thu, Jul 7, 2022 at 9:27 AM Bob Simons - NOAA Federal <
bob.simons@xxxxxxxx> wrote:
> With both netcdf-java v5.5.2 and v5.5.3, the CatalogBuilder no longer
> seems to work with a remoteCatalogService catalog. This code worked before.
>
> Specifically, trying to get a catalog from a remoteCatalogService catalog
> returns a catalog that is null:
> thredds.client.catalog.Catalog catalog = (new
> thredds.client.catalog.builder.CatalogBuilder()).buildFromURI(new
> java.net.URI(
> "
> https://oceanwatch.pfeg.noaa.gov/thredds/remoteCatalogService?catalog=https://thredds1.pfeg.noaa.gov/thredds/catalog/Satellite/MPOC/catalog.xml";));
>
> System.out.println(catalog == null);
>
> That prints "true".
>
> A null is not a useful response. If the URL is invalid for this method,
> then there should be an exception saying so, or more specifically solving
> the problem.
>
> I will solve the problem in my code by using the remote URL directly.
> Since this code used to work, it seems like the best solution is for you to
> have the CatalogBuilder switch to using the remote catalog's url.
>
> Can you please investigate and (if true) fix these problems?
>
> Thank you.
>
