Allow concat_dim_kwargs in THREDDSMergedSource

CHANGELOG.md

@@ -5,6 +5,7 @@
 ### Features
 
 - Allow `xarray_kwargs` for :py:class:`~intake_thredds.THREDDSMergedSource` [#32](https://github.com/NCAR/intake-thredds/pull/32) ([@aaronspring](https://github.com/aaronspring))
+- Allow `concat_dim` for :py:class:`~intake_thredds.THREDDSMergedSource` [#34](https://github.com/NCAR/intake-thredds/pull/34) ([@aaronspring](https://github.com/aaronspring))
 
 ## Intake-thredds v2021.2.17
 

ci/environment-upstream-dev.yml

@@ -17,6 +17,7 @@ dependencies:
   - pytest-xdist
   - tqdm
   - xarray
+  - cfgrib
   - pip:
       - git+https://github.com/intake/intake.git
       - git+https://github.com/Unidata/siphon.git

ci/environment.yml

@@ -19,3 +19,4 @@ dependencies:
   - siphon
   - tqdm
   - xarray
+  - cfgrib

intake_thredds/source.py

@@ -89,14 +89,21 @@ def _open_dataset(self):
                 else:
                     break
             path = self.path[i:]
+            if 'concat_dim' in self.xarray_kwargs:
+                concat_dim = self.xarray_kwargs.pop('concat_dim')
+            else:
+                concat_dim = None
             if self.progressbar:
                 data = [
                     ds(xarray_kwargs=self.xarray_kwargs).to_dask()
                     for ds in tqdm(_match(cat, path), desc='Dataset(s)', ncols=79)
                 ]
             else:
                 data = [ds(xarray_kwargs=self.xarray_kwargs).to_dask() for ds in _match(cat, path)]
-            self._ds = xr.combine_by_coords(data, combine_attrs='override')
+            if concat_dim:
+                self._ds = xr.concat(data, concat_dim)
+            else:
+                self._ds = xr.combine_by_coords(data, combine_attrs='override')
 
 
 def _match(cat, patterns):

tests/test_source.py

@@ -102,7 +102,7 @@ def test_THREDDSMergedSource_simplecache_fails_opendap(THREDDSMergedSource_cat_s
 @pytest.mark.parametrize('driver', ['netcdf', 'opendap'])
 @pytest.mark.parametrize('decode_times', [True, False])
 def test_THREDDSMergedSource_xarray_kwargs(THREDDSMergedSource_cat_short_url, driver, decode_times):
-    """Test xarray_kwargs."""
+    """Test THREDDSMergedSource with xarray_kwargs."""
     ds = intake.open_thredds_merged(
         'https://psl.noaa.gov/thredds/catalog.xml',
         [
@@ -119,3 +119,25 @@ def test_THREDDSMergedSource_xarray_kwargs(THREDDSMergedSource_cat_short_url, dr
         assert 'units' not in ds.time.attrs
     else:
         assert 'units' in ds.time.attrs
+
+
+def test_concat_dim():
+    """Test THREDDSMergedSource with concat_dim. Requires multiple files with same
+    other coords to be concatinated along new dimension specified by concat_dim.
+    Here get two ensemble members initialized 20200831 00:00 at 15.5 days = 372h"""
+    url = 'simplecache::https://www.ncei.noaa.gov/thredds/catalog/model-gefs-003/202008/20200831/catalog.xml'
+    ds = intake.open_thredds_merged(
+        url,
+        ['NCEP gens-a Grid 3 Member-Forecast 1[1-2]*-372 for 2020-08-31 00:00*'],
+        driver='netcdf',
+        xarray_kwargs=dict(
+            engine='cfgrib',
+            concat_dim='number',
+            backend_kwargs=dict(
+                filter_by_keys={'typeOfLevel': 'heightAboveGround', 'shortName': '2t'}
+            ),
+        ),
+    ).to_dask()
+    assert 'number' in ds.dims
+    assert 11 in ds.number
+    assert 12 in ds.number