OK, I get it now.
So the "filename" to open is the exact SUBDATASET_X_NAME string from the
metadata. It seems you can safely omit the quotes (at least in the case of
this file), which is perhaps part of my confusion.
Sorry, that makes sense in hindsight.
Thank you, as ever, for your prompt assis
This seems like a bug. It should still be able to recognize and open
the file without the type hint, whether or not you provide a
sub-dataset suffix.
This is not a bug. This is intended. Subdatasets have a specific syntax.
Valid subdataset names are the ones reported under the SUBDATASET
me
We are dealing with some netCDF files on S3 which have multiple
sub-datasets.
Trying to do a simple gdalinfo using VSI paths.
Note you may need to have AWS_NO_SIGN_REQUEST=YES to access the file below.
This works, and gives a summary dump...
$ gdalinfo /vsis3/noaa-nwm-pds/nwm.20230720/short_ran
