dcherian
diff --git a/‎asv_bench/asv.conf.json
+2 b/‎asv_bench/asv.conf.json
+2
diff --git a/‎asv_bench/benchmarks/groupby.py
+6-4 b/‎asv_bench/benchmarks/groupby.py
+6-4
diff --git a/‎ci/install-upstream-wheels.sh
+2 b/‎ci/install-upstream-wheels.sh
+2
diff --git a/‎ci/requirements/all-but-dask.yml
+1 b/‎ci/requirements/all-but-dask.yml
+1
diff --git a/‎ci/requirements/environment-windows.yml
+1 b/‎ci/requirements/environment-windows.yml
+1
diff --git a/‎ci/requirements/environment.yml
+1 b/‎ci/requirements/environment.yml
+1
diff --git a/‎ci/requirements/min-all-deps.yml
+1 b/‎ci/requirements/min-all-deps.yml
+1
diff --git a/‎doc/whats-new.rst
+5 b/‎doc/whats-new.rst
+5
diff --git a/‎setup.cfg
+1 b/‎setup.cfg
+1
@@ -65,6 +65,8 @@
         "bottleneck": [""],
         "dask": [""],
         "distributed": [""],
+        "flox": [""],
+        "numpy_groupies": [""],
         "sparse": [""]
     },
 
 
@@ -13,6 +13,7 @@ def setup(self, *args, **kwargs):
             {
                 "a": xr.DataArray(np.r_[np.repeat(1, self.n), np.repeat(2, self.n)]),
                 "b": xr.DataArray(np.arange(2 * self.n)),
+                "c": xr.DataArray(np.arange(2 * self.n)),
             }
         )
         self.ds2d = self.ds1d.expand_dims(z=10)
@@ -50,10 +51,11 @@ class GroupByDask(GroupBy):
     def setup(self, *args, **kwargs):
         requires_dask()
         super().setup(**kwargs)
-        self.ds1d = self.ds1d.sel(dim_0=slice(None, None, 2)).chunk({"dim_0": 50})
-        self.ds2d = self.ds2d.sel(dim_0=slice(None, None, 2)).chunk(
-            {"dim_0": 50, "z": 5}
-        )
+
+        self.ds1d = self.ds1d.sel(dim_0=slice(None, None, 2))
+        self.ds1d["c"] = self.ds1d["c"].chunk({"dim_0": 50})
+        self.ds2d = self.ds2d.sel(dim_0=slice(None, None, 2))
+        self.ds2d["c"] = self.ds2d["c"].chunk({"dim_0": 50, "z": 5})
         self.ds1d_mean = self.ds1d.groupby("b").mean()
         self.ds2d_mean = self.ds2d.groupby("b").mean()
 
 
@@ -15,6 +15,7 @@ conda uninstall -y --force \
     pint \
     bottleneck \
     sparse \
+    flox \
     h5netcdf \
     xarray
 # to limit the runtime of Upstream CI
@@ -47,4 +48,5 @@ python -m pip install \
     git+https://github.com/pydata/sparse \
     git+https://github.com/intake/filesystem_spec \
     git+https://github.com/SciTools/nc-time-axis \
+    git+https://github.com/dcherian/flox \
     git+https://github.com/h5netcdf/h5netcdf
@@ -13,6 +13,7 @@ dependencies:
   - cfgrib
   - cftime
   - coveralls
+  - flox
   - h5netcdf
   - h5py
   - hdf5
 
@@ -10,6 +10,7 @@ dependencies:
   - cftime
   - dask-core
   - distributed
+  - flox
   - fsspec!=2021.7.0
   - h5netcdf
   - h5py
 
@@ -12,6 +12,7 @@ dependencies:
   - cftime
   - dask-core
   - distributed
+  - flox
   - fsspec!=2021.7.0
   - h5netcdf
   - h5py
 
@@ -17,6 +17,7 @@ dependencies:
   - coveralls
   - dask-core=2021.04
   - distributed=2021.04
+  - flox=0.5
   - h5netcdf=0.11
   - h5py=3.1
   # hdf5 1.12 conflicts with h5py=3.1
 
@@ -141,6 +141,11 @@ Performance
 - GroupBy binary operations are now vectorized.
   Previously this involved looping over all groups. (:issue:`5804`,:pull:`6160`)
   By `Deepak Cherian <https://github.com/dcherian>`_.
+- Substantially improved GroupBy operations using `flox <https://flox.readthedocs.io/en/latest/>`_.
+  This is auto-enabled when ``flox`` is installed. Use ``xr.set_options(use_flox=False)`` to use
+  the old algorithm. (:issue:`4473`, :issue:`4498`, :issue:`659`, :issue:`2237`, :pull:`271`).
+  By `Deepak Cherian <https://github.com/dcherian>`_,`Anderson Banihirwe <https://github.com/andersy005>`_,
+  `Jimmy Westling <https://github.com/illviljan>`_.
 
 Internal Changes
 ~~~~~~~~~~~~~~~~
 
@@ -98,6 +98,7 @@ accel =
     scipy
     bottleneck
     numbagg
+    flox
 
 parallel =
     dask[complete]