GRIB to NetCDF notebook - cannot replicate

[m9brady]

I'm trying to run through the provided notebook but encountered errors at safely_expand_dims due to a pre-existing valid_time coordinate in the GRIB file I'm working with.

The GRIB in question was downloaded today using cdsapi and can be acquired using the following snippet:

Download sample ERA5L GRIB

import cdsapi

dataset = "reanalysis-era5-land"
request = {
    "product_type": ["reanalysis"],
    "variable": ["10m_u_component_of_wind"],
    "year": ["1993"],
    "month": ["03"],
    "day": [
        "01", "02", "03",
        "04", "05", "06",
        "07", "08", "09",
        "10", "11", "12",
        "13", "14", "15",
        "16", "17", "18",
        "19", "20", "21",
        "22", "23", "24",
        "25", "26", "27",
        "28", "29", "30",
        "31"
    ],
    "time": [
        "00:00", "01:00", "02:00",
        "03:00", "04:00", "05:00",
        "06:00", "07:00", "08:00",
        "09:00", "10:00", "11:00",
        "12:00", "13:00", "14:00",
        "15:00", "16:00", "17:00",
        "18:00", "19:00", "20:00",
        "21:00", "22:00", "23:00"
    ],
    "data_format": "grib",
    "download_format": "unarchived"
}

client = cdsapi.Client()
client.retrieve(dataset, request).download('era5l.1993.03.10m_u_component_of_wind.grib')

the exact error I am seeing is

ValueError: valid_time already exists as coordinate or variable name.

which makes sense because valid_time does exist as a coordinate in the cfgrib-parsed xarray dataset (see below), but is unused for the u10 variable. When the CDS provides a NetCDF version of a GRIB, do these "unused" coordinates get dropped altogether?

<xarray.Dataset> Size: 20GB
Dimensions:     (time: 32, step: 24, latitude: 1801, longitude: 3600)
Coordinates:
    number      int64 8B ...
  * time        (time) datetime64[ns] 256B 1993-02-28 1993-03-01 ... 1993-03-31
  * step        (step) timedelta64[ns] 192B 01:00:00 ... 1 days 00:00:00
    surface     float64 8B ...
  * latitude    (latitude) float64 14kB 90.0 89.9 89.8 ... -89.8 -89.9 -90.0
  * longitude   (longitude) float64 29kB 0.0 0.1 0.2 0.3 ... 359.7 359.8 359.9
    valid_time  (time, step) datetime64[ns] 6kB ...
Data variables:
    u10         (time, step, latitude, longitude) float32 20GB ...
Attributes:
    GRIB_edition:            1
    GRIB_centre:             ecmf
    GRIB_centreDescription:  European Centre for Medium-Range Weather Forecasts
    GRIB_subCentre:          0
    Conventions:             CF-1.7
    institution:             European Centre for Medium-Range Weather Forecasts
    history:                 2025-06-26T16:02 GRIB to CDM+CF via cfgrib-0.9.1...

The history attr provides a clue but I am not sure what is meant by the encode_cf key:

2025-06-26T16:08 GRIB to CDM+CF via cfgrib-0.9.14.1/ecCodes-2.30.2 with {"source": "era5l.1993.03.10m_u_component_of_wind.grib", "filter_by_keys": {}, "encode_cf": ["parameter", "time", "geography", "vertical"]}

As an experiment, I also downloaded the experimental-netcdf version from CDS and see that most of the extra GRIB coordinates are dropped, but a new expver is added that seems to indicate preliminary vs official data releases. This expver isn't commented on in the notebook so I'm even more confused.

I would like to be able to replicate as-closely-as-possible the "Experimental NetCDF" delivery method that the CDS provides but I can't quite seem to figure out what's going on server-side. Is there a public repo containing the various conversion tools used at ECMWF CDS?

Thanks!

[m9brady]

@EddyCMWF any insight? Or is it better to engage by email? We're essentially looking for the GRIB -> NC conversion commands/script that the API uses.