`GroupNotFoundError` with datasets 2025-09+

[ana-sher]

While retrieving datasets for "sentinel-2-l2a" , certain area (will provide in the full code below) and period '2025-07-01/2025-09-30' after filtering out deprecated, got results with a gap (between 2025 09 23 and 2025 07 27), see:

...
'https://objects.eodc.eu:443/e05ab01a9d56408d82ac32d69a5aae2a:202509-s02msil2a/23/products/cpm_v256/S2C_MSIL2A_20250923T095041_N0511_R079_T33UYR_20250923T150014.zarr',
 'https://objects.eodc.eu:443/e05ab01a9d56408d82ac32d69a5aae2a:202507-s02msil2a/27/products/cpm_v256/S2A_MSIL2A_20250727T095101_N0511_R079_T33UYR_20250727T121617.zarr'
...

Datatree for 2025 07 27 zarr file opens ok, but for the datasets after 09 I'm having the following issue:

GroupNotFoundError: No group found in store 'https://objects.eodc.eu:443/e05ab01a9d56408d82ac32d69a5aae2a:202509-s02msil2a/23/products/cpm_v256/S2C_MSIL2A_20250923T095041_N0511_R079_T33UYR_20250923T150014.zarr' at path ''

Full code to reproduce the issue:

from pystac import MediaType, ItemCollection
from pystac_client import Client
import xarray as xr

eopf_stac_api_root_endpoint = "https://stac.core.eopf.eodc.eu/"
collection = "sentinel-2-l2a"
eopf_catalog = Client.open(url=eopf_stac_api_root_endpoint)

def fetch_catalogue_collection(collection_name: str) -> ItemCollection:
    search_result = eopf_catalog.search(
        collections=collection_name,
        bbox=(19.081281771627136, 49.7814303219126, 19.10565768715448, 49.79195922883795),
        datetime='2025-07-01T00:00:00Z/2025-09-30T23:59:59.999999Z',
    )
    return search_result.item_collection()

search_result_s2 = fetch_catalogue_collection(collection)
id_date_collection = [(item.id, item.datetime) for item in search_result_s2 if 'deprecated' not in item.properties or item.properties['deprecated'] == False]

c_sentinel2 = eopf_catalog.get_collection(collection)
items = c_sentinel2.get_items(*[item[0] for item in id_date_collection])
assets = [item.get_assets(media_type=MediaType.ZARR) for item in items]
cloud_storage_urls = [asset['product'].href for asset in assets]

xr.open_datatree(
    cloud_storage_urls[7],
    engine="zarr",
    chunks="auto",
)

[clausmichele]

I think this is related to the removal of the data as planned for the running archive. For instance, this Item is marked as deprecated and only the metadata remains: https://stac.browser.user.eopf.eodc.eu/collections/sentinel-2-l2a/items/S2C_MSIL2A_20250629T110641_N0511_R137_T31UCU_20250629T144518

[senmao]

since this is not related to conversion but item deprecation. I will close this issue now.

[ana-sher]

The code included filtering out for deprecation:
id_date_collection = [(item.id, item.datetime) for item in search_result_s2 if 'deprecated' not in item.properties or item.properties['deprecated'] == False]
Or it wouldn't be sufficient?

[senmao]

Hi @ana-sher, In general, I think filtering out deprecated items should work. But for these items, they don't have deprecated info in their properties and the data on the storage side was removed.

@christophreimer @SerRichard It is an inconsistency issue of deprecating items.