Failing on ECMWF ensemble about gribscan HOT 4 OPEN

I've seen something like this before. I think this was related to variables which occur only in a single message (e.g. a single non-time-varying horizontal field). 2899eac is probably what you need. It's included in the current main, but unfortunately the latest PyPi release is one commit earlier 😬.

• @guidocioni can you try with the latest main version?
• @lkluft do you think we could get some new release?

from gribscan.

I've seen something like this before. I think this was related to variables which occur only in a single message (e.g. a single non-time-varying horizontal field). 2899eac is probably what you need. It's included in the current main, but unfortunately the latest PyPi release is one commit earlier 😬.

• @guidocioni can you try with the latest main version?
• @lkluft do you think we could get some new release?

Let me try again and I'll get back to you.

from gribscan.

Hey, so with the correction that you posted it seems to work, albeit the first step being slower.

For the second step I had to manually create a folder dataset.json, because otherwise it was complaining

Traceback (most recent call last):
  File "/home/ekman/miniconda3/envs/models/bin/gribscan-build", line 8, in <module>
    sys.exit(build_dataset())
  File "/home/ekman/miniconda3/envs/models/lib/python3.10/site-packages/gribscan/tools.py", line 32, in build_dataset
    with open(f"{args.output}/{dataset}.json", "w") as indexfile:
FileNotFoundError: [Errno 2] No such file or directory: 'dataset.json/atm2d.json'

After creating the folder it wrote a json file dataset.json/atm2d.json.
I'm not sure if I'm doing things correctly because I imagine you would have already included a check to see if that folder is existing or not...so it smells like I'm using one of the parameters wrong.

Anyhoe...when I tried to load into xarray

ds = xr.open_zarr("reference::/home/ekman/ssd/guido/ecmwf-ens/dataset.json/atm2d.json", consolidated=False)

Then I get a dataset with collapsed dimensions.

Again, sorry if I'm doing something wrong: zarr is kind of new to me.

from gribscan.

This looks exactly like what I would expect. I also normally create the folder dataset.json or json.dir myself (the one you give with -o) before running the gribscan-build command.

from gribscan.