Add hist.histtype rcParam.

histtype='stepfilled' is significantly faster than the current default ('bar') for large numbers of bins, but changing the default would cause backcompat problems. The 'hist.histtype' rcParam is introduced with the explicit intent of preparing this transition (later making 'stepfilled' (or perhaps 'step') the default and then deprecating the rcParam, to avoid neverending growth of the rcParam list).
anntzer · Oct 11, 2018 · Oct 11, 2018 · Oct 11, 2018 · a770066b3316a5d1e1270e511843e5a1c59e4822 · ImportanceOfBeingErnest
commit a770066b3316a5d1e1270e511843e5a1c59e4822
diff --git a/doc/users/next_whats_new/2018-10-11-AL.rst b/doc/users/next_whats_new/2018-10-11-AL.rst
@@ -0,0 +1,7 @@
+Introduction of :rc:`hist.histtype`
+```````````````````````````````````
+
+The default value of the ``histtype`` argument to `Axes.hist()` can now be
+controlled using :rc:`hist.histtype`.  For large numbers of bins, setting
+this value to 'step' or 'stepfilled' can result in significant performance
+improvements over 'bar', the current default.
diff --git a/lib/matplotlib/axes/_axes.py b/lib/matplotlib/axes/_axes.py
@@ -6259,7 +6259,7 @@ def table(self, **kwargs):
 
     @_preprocess_data(replace_names=["x", 'weights'], label_namer="x")
     def hist(self, x, bins=None, range=None, density=None, weights=None,
-             cumulative=False, bottom=None, histtype='bar', align='mid',
+             cumulative=False, bottom=None, histtype=None, align='mid',
              orientation='vertical', rwidth=None, log=False,
              color=None, label=None, stacked=False, normed=None,
              **kwargs):
@@ -6375,7 +6375,10 @@ def hist(self, x, bins=None, range=None, density=None, weights=None,
             - 'step' generates a lineplot that is by default unfilled.
             - 'stepfilled' generates a lineplot that is by default filled.
 
-            Default is 'bar'.
+            For large numbers of bins, 'step' and 'stepfilled' can be
+            significantly faster than 'bar' and 'barstacked'.
+
+            Default is taken from :rc:`hist.histtype`.
 
         align : {'left', 'mid', 'right'}, optional
             Controls how the histogram is plotted.
@@ -6471,6 +6474,8 @@ def hist(self, x, bins=None, range=None, density=None, weights=None,
 
         if bins is None:
             bins = rcParams['hist.bins']
+        if histtype is None:
+            histtype = rcParams['hist.histtype']
 
         # Validate string inputs here so we don't have to clutter
         # subsequent code.

diff --git a/lib/matplotlib/pyplot.py b/lib/matplotlib/pyplot.py
@@ -2625,7 +2625,7 @@ def hexbin(
 @docstring.copy_dedent(Axes.hist)
 def hist(
         x, bins=None, range=None, density=None, weights=None,
-        cumulative=False, bottom=None, histtype='bar', align='mid',
+        cumulative=False, bottom=None, histtype=None, align='mid',
         orientation='vertical', rwidth=None, log=False, color=None,
         label=None, stacked=False, normed=None, *, data=None,
         **kwargs):

diff --git a/lib/matplotlib/rcsetup.py b/lib/matplotlib/rcsetup.py
@@ -1034,6 +1034,10 @@ def _validate_linestyle(ls):
 
     ## Histogram properties
     'hist.bins': [10, validate_hist_bins],
+    'hist.histtype': [
+        'bar',
+        ValidateInStrings(
+            'histtype', ['bar', 'barstacked', 'step', 'stepfilled'])],
 
     ## Boxplot properties
     'boxplot.notch': [False, validate_bool],

diff --git a/matplotlibrc.template b/matplotlibrc.template
@@ -481,6 +481,7 @@
 #hist.bins : 10                   ## The default number of histogram bins.
                                   ## If Numpy 1.11 or later is
                                   ## installed, may also be `auto`
+#hist.histtype : bar              ## The default histtype.
 
 #### SCATTER PLOTS
 #scatter.marker : o               ## The default marker type for scatter plots.