Fixed KDE Plot to drop the missing values (pandas-dev#14820)

BUG: Fixed KDE plot to ignore missing values

 closes pandas-dev#14821

* fixed kde plot to ignore the missing values
* added comment to elaborate the changes made
* added a release note in whatsnew/0.19.2
* added test to check for  missing values and cleaned up whatsnew doc
* added comment to refer the issue
* modified to fit lint checks
* replaced ._xorig with .get_xdata()
(cherry picked from commit 033d345)

Author: keshavramaswamy

doc/source/whatsnew/v0.19.2.txt

@@ -88,4 +88,6 @@ Bug Fixes
 
 - Explicit check in ``to_stata`` and ``StataWriter`` for out-of-range values when writing doubles (:issue:`14618`)
 
+- Bug in ``.plot(kind='kde')`` which did not drop missing values to generate the KDE Plot, instead generating an empty plot. (:issue:`14821`)
+
 - Bug in ``unstack()`` if called with a list of column(s) as an argument, regardless of the dtypes of all columns, they get coerced to ``object`` (:issue:`11847`)

pandas/tests/plotting/test_series.py

@@ -569,7 +569,11 @@ def test_kde_missing_vals(self):
         _skip_if_no_scipy_gaussian_kde()
         s = Series(np.random.uniform(size=50))
         s[0] = np.nan
-        _check_plot_works(s.plot.kde)
+        axes = _check_plot_works(s.plot.kde)
+        # check if the values have any missing values
+        # GH14821
+        self.assertTrue(any(~np.isnan(axes.lines[0].get_xdata())),
+                        msg='Missing Values not dropped')
 
     @slow
     def test_hist_kwargs(self):

pandas/tools/plotting.py

@@ -2136,9 +2136,10 @@ def _args_adjust(self):
 
     def _get_ind(self, y):
         if self.ind is None:
-            sample_range = max(y) - min(y)
-            ind = np.linspace(min(y) - 0.5 * sample_range,
-                              max(y) + 0.5 * sample_range, 1000)
+            # np.nanmax() and np.nanmin() ignores the missing values
+            sample_range = np.nanmax(y) - np.nanmin(y)
+            ind = np.linspace(np.nanmin(y) - 0.5 * sample_range,
+                              np.nanmax(y) + 0.5 * sample_range, 1000)
         else:
             ind = self.ind
         return ind