From 617628ff10840a438f7616605636161cb0ebbee1 Mon Sep 17 00:00:00 2001
From: Illviljan <14371165+Illviljan@users.noreply.github.com>
Date: Tue, 24 Jan 2023 19:48:09 +0100
Subject: [PATCH] Add benchmarks for to_dataframe and to_dask_dataframe

---
 asv_bench/benchmarks/pandas.py | 37 +++++++++++++++++++++++++++++++++-
 1 file changed, 36 insertions(+), 1 deletion(-)

diff --git a/asv_bench/benchmarks/pandas.py b/asv_bench/benchmarks/pandas.py
index 8aaa515d417..2a296ecc4d0 100644
--- a/asv_bench/benchmarks/pandas.py
+++ b/asv_bench/benchmarks/pandas.py
@@ -3,7 +3,7 @@
 
 import xarray as xr
 
-from . import parameterized
+from . import parameterized, requires_dask
 
 
 class MultiIndexSeries:
@@ -24,3 +24,38 @@ def setup(self, dtype, subset):
     @parameterized(["dtype", "subset"], ([int, float], [True, False]))
     def time_from_series(self, dtype, subset):
         xr.DataArray.from_series(self.series)
+
+
+class ToDataFrame:
+    def setup(self, *args, **kwargs):
+        xp = kwargs.get("xp", np)
+        random_kws = kwargs.get("random_kws", {})
+        method = kwargs.get("method", "to_dataframe")
+
+        dim1 = 10_000
+        dim2 = 10_000
+        ds = xr.Dataset(
+            {
+                "x": xr.DataArray(
+                    data=xp.random.random((dim1, dim2), **random_kws),
+                    dims=["dim1", "dim2"],
+                    coords={"dim1": np.arange(0, dim1), "dim2": np.arange(0, dim2)},
+                )
+            }
+        )
+        self.to_frame = getattr(ds, method)
+
+    def time_to_dataframe(self):
+        self.to_frame()
+
+    def peakmem_to_dataframe(self):
+        self.to_frame()
+
+
+class ToDataFrameDask(ToDataFrame):
+    def setup(self, *args, **kwargs):
+        requires_dask()
+
+        import dask.array as da
+
+        super().setup(xp=da, random_kws=dict(chunks=5000), method="to_dask_dataframe")