plot.contour: Don't make cmap if colors is a single color. (#2453)

By default, matplotlib draw dashed negative contours for a single color.
We lost this feature by manually specifying cmap everytime.

doc/whats-new.rst

@@ -68,6 +68,10 @@ Bug fixes
 - ``xarray.DataArray.std()`` now correctly accepts ``ddof`` keyword argument.
   (:issue:`2240`)
   By `Keisuke Fujii <https://github.com/fujiisoup>`_.
+- Restore matplotlib's default of plotting dashed negative contours when
+  a single color is passed to ``DataArray.contour()`` e.g. ``colors='k'``.
+  By `Deepak Cherian <https://github.com/dcherian>`_.
+
 
 .. _whats-new.0.10.9:
 

xarray/plot/plot.py

@@ -737,6 +737,11 @@ def newplotfunc(darray, x=None, y=None, figsize=None, size=None,
             # pcolormesh
             kwargs['extend'] = cmap_params['extend']
             kwargs['levels'] = cmap_params['levels']
+            # if colors == a single color, matplotlib draws dashed negative
+            # contours. we lose this feature if we pass cmap and not colors
+            if isinstance(colors, basestring):
+                cmap_params['cmap'] = None
+                kwargs['colors'] = colors
 
         if 'pcolormesh' == plotfunc.__name__:
             kwargs['infer_intervals'] = infer_intervals

xarray/tests/test_plot.py

@@ -1140,9 +1140,9 @@ def test_colors(self):
         def _color_as_tuple(c):
             return tuple(c[:3])
 
+        # with single color, we don't want rgb array
         artist = self.plotmethod(colors='k')
-        assert _color_as_tuple(artist.cmap.colors[0]) == \
-            (0.0, 0.0, 0.0)
+        assert artist.cmap.colors[0] == 'k'
 
         artist = self.plotmethod(colors=['k', 'b'])
         assert _color_as_tuple(artist.cmap.colors[1]) == \