Remove nans early in distplot to avoid reference rule errors

seaborn/distributions.py

@@ -179,9 +179,7 @@ def distplot(a, bins=None, hist=True, kde=True, rug=False, fit=None,
         a = a.squeeze()
 
     # Drop null values from array
-    mask = np.isnan(a)
-    if mask.any():
-        a = a[~mask]
+    a = remove_na(a)
 
     # Decide if the hist is normed
     norm_hist = norm_hist or kde or (fit is not None)

seaborn/tests/test_distributions.py

@@ -16,6 +16,26 @@
     _no_statsmodels = True
 
 
+class TestDistPlot(object):
+
+    def test_distplot_with_nans(self):
+
+        f, (ax1, ax2) = plt.subplots(2)
+        x = np.random.randn(100)
+        x_null = np.append(x, [np.nan])
+
+        dist.distplot(x, ax=ax1)
+        dist.distplot(x_null, ax=ax2)
+
+        line1 = ax1.lines[0]
+        line2 = ax2.lines[0]
+        assert np.array_equal(line1.get_xydata(), line2.get_xydata())
+
+        for bar1, bar2 in zip(ax1.patches, ax2.patches):
+            assert bar1.get_xy() == bar2.get_xy()
+            assert bar1.get_height() == bar2.get_height()
+
+
 class TestKDE(object):
 
     rs = np.random.RandomState(0)
@@ -110,7 +130,7 @@ def test_kde_singular(self):
         assert not line.get_xydata().size
 
     @pytest.mark.parametrize("cumulative", [True, False])
-    def test_kdeplot_missing(self, cumulative):
+    def test_kdeplot_with_nans(self, cumulative):
 
         x_missing = np.append(self.x, [np.nan, np.nan])