Don't fail when plotting Series/DataFrame with no row

When Series or DataFrames was empty (no cells) after removing columns
without a suitable type (.select_dtypes), pandas was throwing a "no
numeric data to plot" exception.

This can happen:

1) either because there is no column with a suitable type;

2) or if there is no row.

Raising an exception in the first case makes sense but we should
probably avoid throwing an exception in the second case.

Author: Gabriel Corona

doc/source/whatsnew/v1.0.0.rst

@@ -168,6 +168,7 @@ Plotting
 
 - Bug in :meth:`Series.plot` not able to plot boolean values (:issue:`23719`)
 -
+- Bug in :meth:`DataFrame.plot` not able to plot when no rows (:issue:`27758`)
 - Bug in :meth:`DataFrame.plot` producing incorrect legend markers when plotting multiple series on the same axis (:issue:`18222`)
 - Bug in :meth:`DataFrame.plot` when ``kind='box'`` and data contains datetime or timedelta data. These types are now automatically dropped (:issue:`22799`)
 - Bug in :meth:`DataFrame.plot.line` and :meth:`DataFrame.plot.area` produce wrong xlim in x-axis (:issue:`27686`, :issue:`25160`, :issue:`24784`)

pandas/plotting/_matplotlib/core.py

@@ -418,7 +418,7 @@ def _compute_plot_data(self):
         numeric_data = data.select_dtypes(include=include_type, exclude=exclude_type)
 
         try:
-            is_empty = numeric_data.empty
+            is_empty = numeric_data.columns.empty
         except AttributeError:
             is_empty = not len(numeric_data)
 

pandas/tests/plotting/test_frame.py

@@ -3229,6 +3229,21 @@ def test_subplots_sharex_false(self):
         tm.assert_numpy_array_equal(axs[0].get_xticks(), expected_ax1)
         tm.assert_numpy_array_equal(axs[1].get_xticks(), expected_ax2)
 
+    def test_plot_no_rows(self):
+        # GH 27758
+        df = pd.DataFrame(columns=["foo"], index=pd.Index(data=[], dtype=int), dtype=int)
+        assert df.empty
+        ax = df.plot()
+        assert len(ax.get_lines()) == 1
+        line = ax.get_lines()[0]
+        assert len(line.get_xdata()) == 0
+        assert len(line.get_ydata()) == 0
+
+    def test_plot_no_numeric_data(self):
+        df = pd.DataFrame(["a", "b", "c"])
+        with pytest.raises(TypeError):
+            df.plot()
+
 
 def _generate_4_axes_via_gridspec():
     import matplotlib.pyplot as plt