Use cumsum from flox

* use cumsum from flox

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update groupby.py

* Update groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update groupby.py

* Update groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* use apply_ufunc for dataset and dataarray handling

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* sync protocols with each other

* typing

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* add dataset and version requirement

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update _aggregations.py

* Update xarray/core/groupby.py

Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>

* Update groupby.py

* Update groupby.py

* Update groupby.py

* Update generate_aggregations.py

* Renove workaround in test

* Update _aggregations.py

* Update _aggregations.py

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* clean ups

* Add expected groups, add options

* Update groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* expeced_groups not supported in groupby_scan

* Update _aggregations.py

* Update _aggregations.py

* Update generate_aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update generate_aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update test_groupby.py

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update test_groupby.py

* Update generate_aggregations.py

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update generate_aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Update test_groupby.py

* Update generate_aggregations.py

* Update _aggregations.py

* Update _aggregations.py

* Datatree should use Dataset.func rather than Dataset.reduce(func)

* docstsrings for datatree can now be fixed

* Update _aggregations.py

* Update test_dataset.py

* Update test_dataset.py

* Update generate_aggregations.py

* Update _aggregations.py

* Update generate_aggregations.py

* Update generate_aggregations.py

* Update _aggregations.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* mapping in collections

* add coords

* test ds.cumsum with expected outcomes

* Update test_dataset.py

* Update test_dataset.py

* Update test_dataset.py

* Add use_lazy_group_idx to parametrize

* Update test_groupby.py

* cumusum in flox requires 0.10.5

* Update test_groupby.py

* Update whats-new.rst

* Update whats-new.rst

* Update whats-new.rst

* Add ASV test

* Motivate use Dataset.func further

* test refactor

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Revert "[pre-commit.ci] auto fixes from pre-commit.com hooks"

This reverts commit eff5561.

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* nd-array tests and different dims

* Update test_groupby.py

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* more tests

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update test_groupby.py

* Update test_groupby.py

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* Update test_groupby.py

* ns to us

* Update whats-new.rst

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Deepak Cherian <dcherian@users.noreply.github.com>

Author: 3 people

asv_bench/benchmarks/groupby.py

@@ -28,15 +28,17 @@ def time_init(self, ndim):
         getattr(self, f"ds{ndim}d").groupby("b")
 
     @parameterized(
-        ["method", "ndim", "use_flox"], [("sum", "mean"), (1, 2), (True, False)]
+        ["method", "ndim", "use_flox"],
+        [("sum", "mean", "cumsum"), (1, 2), (True, False)],
     )
     def time_agg_small_num_groups(self, method, ndim, use_flox):
         ds = getattr(self, f"ds{ndim}d")
         with xr.set_options(use_flox=use_flox):
             getattr(ds.groupby("a"), method)().compute()
 
     @parameterized(
-        ["method", "ndim", "use_flox"], [("sum", "mean"), (1, 2), (True, False)]
+        ["method", "ndim", "use_flox"],
+        [("sum", "mean", "cumsum"), (1, 2), (True, False)],
     )
     def time_agg_large_num_groups(self, method, ndim, use_flox):
         ds = getattr(self, f"ds{ndim}d")

doc/whats-new.rst

@@ -125,6 +125,13 @@ Documentation
 - Fix code blocks on "how to create custom index" doc page (:pull:`11255`).
   By `Nick Hodgskin <https://github.com/VeckoTheGecko>`_.
 
+Performance
+~~~~~~~~~~~
+
+- Groupby cumsum can now be accelerated with flox. Coordinates are now retained
+  as well. (:issue:`6528`, :pull:`10987`)
+  By `Jimmy Westling <https://github.com/illviljan>`_.
+
 Internal Changes
 ~~~~~~~~~~~~~~~~
 
@@ -345,6 +352,9 @@ Performance
 
 - Add a fastpath to the backend plugin system for standard engines (:issue:`10178`, :pull:`10937`).
   By `Sam Levang <https://github.com/slevang>`_.
+- Groupby cumsum can now be accelerated with flox. Coordinates are now retained
+  as well. (:issue:`6528`, :pull:`10987`)
+  By `Jimmy Westling <https://github.com/illviljan>`_.
 - Optimize :py:class:`~xarray.coding.variables.CFMaskCoder` decoder (:pull:`11105`).
   By `Deepak Cherian <https://github.com/dcherian>`_.