no query params for h5netcdf

Author: ianhi

xarray/backends/h5netcdf_.py

@@ -462,19 +462,13 @@ class H5netcdfBackendEntrypoint(BackendEntrypoint):
     supports_groups = True
 
     def guess_can_open(self, filename_or_obj: T_PathFileOrDataStore) -> bool:
-        from xarray.core.utils import is_remote_uri, strip_uri_params
-
         filename_or_obj = _normalize_filename_or_obj(filename_or_obj)
         magic_number = try_read_magic_number_from_file_or_path(filename_or_obj)
         if magic_number is not None:
             return magic_number.startswith(b"\211HDF\r\n\032\n")
 
         if isinstance(filename_or_obj, str | os.PathLike):
-            path = str(filename_or_obj)
-            # For remote URIs, strip query parameters and fragments before checking extension
-            if isinstance(filename_or_obj, str) and is_remote_uri(path):
-                path = strip_uri_params(path)
-            _, ext = os.path.splitext(path)
+            _, ext = os.path.splitext(str(filename_or_obj))
             return ext in {".nc", ".nc4", ".cdf"}
 
         return False

xarray/tests/test_backends.py

@@ -7281,27 +7281,29 @@ def test_remote_url_backend_auto_detection() -> None:
         ("https://example.com/store.zarr", "zarr"),
         ("http://example.com/data.zarr/", "zarr"),
         ("s3://bucket/path/to/data.zarr", "zarr"),
-        # Remote netCDF URLs (non-DAP) - h5netcdf wins (first in order)
+        # Remote netCDF URLs (non-DAP) - h5netcdf wins (first in order, no query params)
         ("https://example.com/file.nc", "h5netcdf"),
         ("http://example.com/data.nc4", "h5netcdf"),
         ("https://example.com/test.cdf", "h5netcdf"),
-        ("https://example.com/data.nc?var=temperature&time=0", "h5netcdf"),
-        # DAP URLs with query parameters - h5netcdf wins (has .nc4 ext, first in order)
+        ("s3://bucket/path/to/data.nc", "h5netcdf"),
+        # Remote netCDF URLs with query params - netcdf4 wins
+        # Note: Query params are typically indicative of DAP URLs (e.g., OPeNDAP constraint expressions),
+        # so we prefer netcdf4 (which has DAP support) over h5netcdf (which doesn't)
+        ("https://example.com/data.nc?var=temperature&time=0", "netcdf4"),
         (
             "http://test.opendap.org/opendap/dap4/StaggeredGrid.nc4?dap4.ce=/time[0:1:0]",
-            "h5netcdf",
+            "netcdf4",
         ),
         # DAP URLs without extensions - pydap wins
         ("dap2://opendap.earthdata.nasa.gov/collections/dataset", "pydap"),
         ("dap4://opendap.earthdata.nasa.gov/collections/dataset", "pydap"),
         ("DAP2://example.com/dataset", "pydap"),  # uppercase scheme
         ("DAP4://example.com/dataset", "pydap"),  # uppercase scheme
         ("https://example.com/services/DAP2/dataset", "pydap"),  # uppercase in path
-        # DAP URLs with .nc extensions - h5netcdf wins (first in order)
+        # DAP URLs with .nc extensions (no query params) - h5netcdf wins (first in order)
         ("http://test.opendap.org/opendap/dap4/StaggeredGrid.nc4", "h5netcdf"),
         ("https://example.com/DAP4/data.nc", "h5netcdf"),
         ("http://example.com/data/Dap4/file.nc", "h5netcdf"),
-        ("s3://bucket/path/to/data.nc", "h5netcdf"),
     ]
 
     for url, expected_backend in test_cases: