Add a simple nbytes representation in DataArrays and Dataset repr (opt-in) (#8702)

* Add nbytes to repr

* Disable by default

* What's new

* Specify dtype explictly as Windows seems to default on int32

* Use int16

* PR comments: Removed global option + size format in repr header

* Remove width justified

* Remove problematic backslashes

* pytest-accept

* Revert "Remove problematic backslashes"

This reverts commit 85f6ee4.

* Fix missing backslash escape

* Fix ellipsis in excess leading to excess last line

* Update reprs in tests

* Fix tests

* Try windows specific expected outputs

* Conditional Windows testing

* Fix indent

* Flaky test + fix windows

Author: etienneschalk

doc/whats-new.rst

@@ -23,6 +23,9 @@ v2024.02.0 (unreleased)
 New Features
 ~~~~~~~~~~~~
 
+- Added a simple `nbytes` representation in DataArrays and Dataset `repr`.
+  (:issue:`8690`, :pull:`8702`).
+  By `Etienne Schalk <https://github.com/etienneschalk>`_.
 - Allow negative frequency strings (e.g. ``"-1YE"``). These strings are for example used
   in :py:func:`date_range`,  and :py:func:`cftime_range` (:pull:`8651`).
   By `Mathias Hauser <https://github.com/mathause>`_.

xarray/backends/api.py

@@ -1431,12 +1431,12 @@ def save_mfdataset(
     ...     coords={"time": pd.date_range("2010-01-01", freq="ME", periods=48)},
     ... )
     >>> ds
-    <xarray.Dataset>
+    <xarray.Dataset> Size: 768B
     Dimensions:  (time: 48)
     Coordinates:
-      * time     (time) datetime64[ns] 2010-01-31 2010-02-28 ... 2013-12-31
+      * time     (time) datetime64[ns] 384B 2010-01-31 2010-02-28 ... 2013-12-31
     Data variables:
-        a        (time) float64 0.0 0.02128 0.04255 0.06383 ... 0.9574 0.9787 1.0
+        a        (time) float64 384B 0.0 0.02128 0.04255 ... 0.9574 0.9787 1.0
     >>> years, datasets = zip(*ds.groupby("time.year"))
     >>> paths = [f"{y}.nc" for y in years]
     >>> xr.save_mfdataset(datasets, paths)

xarray/coding/cftimeindex.py

@@ -383,30 +383,30 @@ def _partial_date_slice(self, resolution, parsed):
         ...     dims=["time"],
         ... )
         >>> da.sel(time="2001-01-01")
-        <xarray.DataArray (time: 1)>
+        <xarray.DataArray (time: 1)> Size: 8B
         array([1])
         Coordinates:
-          * time     (time) object 2001-01-01 00:00:00
+          * time     (time) object 8B 2001-01-01 00:00:00
         >>> da = xr.DataArray(
         ...     [1, 2],
         ...     coords=[[pd.Timestamp(2001, 1, 1), pd.Timestamp(2001, 2, 1)]],
         ...     dims=["time"],
         ... )
         >>> da.sel(time="2001-01-01")
-        <xarray.DataArray ()>
+        <xarray.DataArray ()> Size: 8B
         array(1)
         Coordinates:
-            time     datetime64[ns] 2001-01-01
+            time     datetime64[ns] 8B 2001-01-01
         >>> da = xr.DataArray(
         ...     [1, 2],
         ...     coords=[[pd.Timestamp(2001, 1, 1, 1), pd.Timestamp(2001, 2, 1)]],
         ...     dims=["time"],
         ... )
         >>> da.sel(time="2001-01-01")
-        <xarray.DataArray (time: 1)>
+        <xarray.DataArray (time: 1)> Size: 8B
         array([1])
         Coordinates:
-          * time     (time) datetime64[ns] 2001-01-01T01:00:00
+          * time     (time) datetime64[ns] 8B 2001-01-01T01:00:00
         """
         start, end = _parsed_string_to_bounds(self.date_type, resolution, parsed)