dataframe bar plot can now accept width and pos keywords for more flexible alignment

Author: sinhrks

doc/source/release.rst

@@ -140,6 +140,9 @@ API Changes
   the index, rather than requiring a list of tuple (:issue:`4370`)
 
 - Fix a bug where invalid eval/query operations would blow the stack (:issue:`5198`)
+- Following keywords are now acceptable for :meth:`DataFrame.plot(kind='bar')` and :meth:`DataFrame.plot(kind='barh')`.
+  - `width`: Specify the bar width. In previous versions, static value 0.5 was passed to matplotlib and it cannot be overwritten.
+  - `position`: Specify relative alignments for bar plot layout. From 0 (left/bottom-end) to 1(right/top-end). Default is 0.5 (center). (:issue:`6604`)
 
 Deprecations
 ~~~~~~~~~~~~

doc/source/v0.14.0.txt

@@ -178,6 +178,10 @@ These are out-of-bounds selections
   ``FutureWarning`` is raised  to alert that the old ``rows`` and ``cols`` arguments
   will not be supported in a future release (:issue:`5505`)
 
+- Following keywords are now acceptable for :meth:`DataFrame.plot(kind='bar')` and :meth:`DataFrame.plot(kind='barh')`.
+  - `width`: Specify the bar width. In previous versions, static value 0.5 was passed to matplotlib and it cannot be overwritten.
+  - `position`: Specify relative alignments for bar plot layout. From 0 (left/bottom-end) to 1(right/top-end). Default is 0.5 (center). (:issue:`6604`)
+
 
 MultiIndexing Using Slicers
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~

pandas/tests/test_graphics.py

@@ -71,87 +71,6 @@ def test_plot_figsize_and_title(self):
         assert_array_equal(np.round(ax.figure.get_size_inches()),
                            np.array((16., 8.)))
 
-    @slow
-    def test_bar_colors(self):
-        import matplotlib.pyplot as plt
-        import matplotlib.colors as colors
-
-        default_colors = plt.rcParams.get('axes.color_cycle')
-        custom_colors = 'rgcby'
-
-        df = DataFrame(randn(5, 5))
-        ax = df.plot(kind='bar')
-
-        rects = ax.patches
-
-        conv = colors.colorConverter
-        for i, rect in enumerate(rects[::5]):
-            xp = conv.to_rgba(default_colors[i % len(default_colors)])
-            rs = rect.get_facecolor()
-            self.assertEqual(xp, rs)
-
-        tm.close()
-
-        ax = df.plot(kind='bar', color=custom_colors)
-
-        rects = ax.patches
-
-        conv = colors.colorConverter
-        for i, rect in enumerate(rects[::5]):
-            xp = conv.to_rgba(custom_colors[i])
-            rs = rect.get_facecolor()
-            self.assertEqual(xp, rs)
-
-        tm.close()
-        from matplotlib import cm
-
-        # Test str -> colormap functionality
-        ax = df.plot(kind='bar', colormap='jet')
-
-        rects = ax.patches
-
-        rgba_colors = lmap(cm.jet, np.linspace(0, 1, 5))
-        for i, rect in enumerate(rects[::5]):
-            xp = rgba_colors[i]
-            rs = rect.get_facecolor()
-            self.assertEqual(xp, rs)
-
-        tm.close()
-
-        # Test colormap functionality
-        ax = df.plot(kind='bar', colormap=cm.jet)
-
-        rects = ax.patches
-
-        rgba_colors = lmap(cm.jet, np.linspace(0, 1, 5))
-        for i, rect in enumerate(rects[::5]):
-            xp = rgba_colors[i]
-            rs = rect.get_facecolor()
-            self.assertEqual(xp, rs)
-
-        tm.close()
-        df.ix[:, [0]].plot(kind='bar', color='DodgerBlue')
-
-    @slow
-    def test_bar_linewidth(self):
-        df = DataFrame(randn(5, 5))
-
-        # regular
-        ax = df.plot(kind='bar', linewidth=2)
-        for r in ax.patches:
-            self.assertEqual(r.get_linewidth(), 2)
-
-        # stacked
-        ax = df.plot(kind='bar', stacked=True, linewidth=2)
-        for r in ax.patches:
-            self.assertEqual(r.get_linewidth(), 2)
-
-        # subplots
-        axes = df.plot(kind='bar', linewidth=2, subplots=True)
-        for ax in axes:
-            for r in ax.patches:
-                self.assertEqual(r.get_linewidth(), 2)
-
     @slow
     def test_bar_log(self):
         expected = np.array([1., 10., 100., 1000.])
@@ -574,6 +493,170 @@ def test_subplots(self):
             [self.assert_(label.get_visible())
              for label in ax.get_yticklabels()]
 
+    @slow
+    def test_bar_colors(self):
+        import matplotlib.pyplot as plt
+        import matplotlib.colors as colors
+
+        default_colors = plt.rcParams.get('axes.color_cycle')
+        custom_colors = 'rgcby'
+
+        df = DataFrame(randn(5, 5))
+        ax = df.plot(kind='bar')
+
+        rects = ax.patches
+
+        conv = colors.colorConverter
+        for i, rect in enumerate(rects[::5]):
+            xp = conv.to_rgba(default_colors[i % len(default_colors)])
+            rs = rect.get_facecolor()
+            self.assertEqual(xp, rs)
+
+        tm.close()
+
+        ax = df.plot(kind='bar', color=custom_colors)
+
+        rects = ax.patches
+
+        conv = colors.colorConverter
+        for i, rect in enumerate(rects[::5]):
+            xp = conv.to_rgba(custom_colors[i])
+            rs = rect.get_facecolor()
+            self.assertEqual(xp, rs)
+
+        tm.close()
+        from matplotlib import cm
+
+        # Test str -> colormap functionality
+        ax = df.plot(kind='bar', colormap='jet')
+
+        rects = ax.patches
+
+        rgba_colors = lmap(cm.jet, np.linspace(0, 1, 5))
+        for i, rect in enumerate(rects[::5]):
+            xp = rgba_colors[i]
+            rs = rect.get_facecolor()
+            self.assertEqual(xp, rs)
+
+        tm.close()
+
+        # Test colormap functionality
+        ax = df.plot(kind='bar', colormap=cm.jet)
+
+        rects = ax.patches
+
+        rgba_colors = lmap(cm.jet, np.linspace(0, 1, 5))
+        for i, rect in enumerate(rects[::5]):
+            xp = rgba_colors[i]
+            rs = rect.get_facecolor()
+            self.assertEqual(xp, rs)
+
+        tm.close()
+        df.ix[:, [0]].plot(kind='bar', color='DodgerBlue')
+
+    @slow
+    def test_bar_linewidth(self):
+        df = DataFrame(randn(5, 5))
+
+        # regular
+        ax = df.plot(kind='bar', linewidth=2)
+        for r in ax.patches:
+            self.assertEqual(r.get_linewidth(), 2)
+
+        # stacked
+        ax = df.plot(kind='bar', stacked=True, linewidth=2)
+        for r in ax.patches:
+            self.assertEqual(r.get_linewidth(), 2)
+
+        # subplots
+        axes = df.plot(kind='bar', linewidth=2, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(r.get_linewidth(), 2)
+
+    @slow
+    def test_bar_barwidth(self):
+        df = DataFrame(randn(5, 5))
+        
+        width = 0.9
+        
+        # regular
+        ax = df.plot(kind='bar', width=width)
+        for r in ax.patches:
+            self.assertEqual(r.get_width(), width / len(df.columns))
+
+        # stacked
+        ax = df.plot(kind='bar', stacked=True, width=width)
+        for r in ax.patches:
+            self.assertEqual(r.get_width(), width)
+
+        # horizontal regular
+        ax = df.plot(kind='barh', width=width)
+        for r in ax.patches:
+            self.assertEqual(r.get_height(), width / len(df.columns))
+
+        # horizontal stacked
+        ax = df.plot(kind='barh', stacked=True, width=width)
+        for r in ax.patches:
+            self.assertEqual(r.get_height(), width)
+
+        # subplots
+        axes = df.plot(kind='bar', width=width, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(r.get_width(), width)
+
+        # horizontal subplots
+        axes = df.plot(kind='barh', width=width, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(r.get_height(), width)
+
+    @slow
+    def test_bar_barwidth_position(self):
+        df = DataFrame(randn(5, 5))
+
+        width = 0.9
+        position = 0.2
+        
+        # regular
+        ax = df.plot(kind='bar', width=width, position=position)
+        p = ax.patches[0]
+        self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                         p.get_x() + p.get_width() * position * len(df.columns))
+
+        # stacked
+        ax = df.plot(kind='bar', stacked=True, width=width, position=position)
+        p = ax.patches[0]
+        self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                         p.get_x() + p.get_width() * position)
+
+        # horizontal regular
+        ax = df.plot(kind='barh', width=width, position=position)
+        p = ax.patches[0]
+        self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         p.get_y() + p.get_height() * position * len(df.columns))
+
+        # horizontal stacked
+        ax = df.plot(kind='barh', stacked=True, width=width, position=position)
+        p = ax.patches[0]
+        self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         p.get_y() + p.get_height() * position)
+
+        # subplots
+        axes = df.plot(kind='bar', width=width, position=position, subplots=True)
+        for ax in axes:
+            p = ax.patches[0]
+            self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                             p.get_x() + p.get_width() * position)
+
+        # horizontal subplots
+        axes = df.plot(kind='barh', width=width, position=position, subplots=True)
+        for ax in axes:
+            p = ax.patches[0]
+            self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                             p.get_y() + p.get_height() * position)
+
     @slow
     def test_plot_scatter(self):
         from matplotlib.pylab import close
@@ -616,11 +699,61 @@ def test_bar_stacked_center(self):
         self.assertEqual(ax.xaxis.get_ticklocs()[0],
                          ax.patches[0].get_x() + ax.patches[0].get_width() / 2)
 
+        ax = df.plot(kind='bar', stacked='True', width=0.9, grid=True)
+        self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                         ax.patches[0].get_x() + ax.patches[0].get_width() / 2)
+
+        ax = df.plot(kind='barh', stacked='True', grid=True)
+        self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         ax.patches[0].get_y() + ax.patches[0].get_height() / 2)
+
+        ax = df.plot(kind='barh', stacked='True', width=0.9, grid=True)
+        self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         ax.patches[0].get_y() + ax.patches[0].get_height() / 2)
+
     def test_bar_center(self):
         df = DataFrame({'A': [3] * 5, 'B': lrange(5)}, index=lrange(5))
         ax = df.plot(kind='bar', grid=True)
         self.assertEqual(ax.xaxis.get_ticklocs()[0],
                          ax.patches[0].get_x() + ax.patches[0].get_width())
+        
+        ax = df.plot(kind='bar', width=0.9, grid=True)
+        self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                         ax.patches[0].get_x() + ax.patches[0].get_width())
+
+        ax = df.plot(kind='barh', grid=True)
+        self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         ax.patches[0].get_y() + ax.patches[0].get_height())
+        
+        ax = df.plot(kind='barh', width=0.9, grid=True)
+        self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         ax.patches[0].get_y() + ax.patches[0].get_height())
+
+    def test_bar_subplots_center(self):
+        df = DataFrame({'A': [3] * 5, 'B': lrange(5)}, index=lrange(5))
+        axes = df.plot(kind='bar', grid=True, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                         ax.patches[0].get_x() + ax.patches[0].get_width() / 2)
+
+        axes = df.plot(kind='bar', width=0.9, grid=True, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(ax.xaxis.get_ticklocs()[0],
+                         ax.patches[0].get_x() + ax.patches[0].get_width() / 2)
+
+        axes = df.plot(kind='barh', grid=True, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         ax.patches[0].get_y() + ax.patches[0].get_height() / 2)
+
+        axes = df.plot(kind='barh', width=0.9, grid=True, subplots=True)
+        for ax in axes:
+            for r in ax.patches:
+                self.assertEqual(ax.yaxis.get_ticklocs()[0],
+                         ax.patches[0].get_y() + ax.patches[0].get_height() / 2)
 
     @slow
     def test_bar_log_no_subplots(self):

pandas/tools/plotting.py

@@ -1672,15 +1672,20 @@ class BarPlot(MPLPlot):
     def __init__(self, data, **kwargs):
         self.mark_right = kwargs.pop('mark_right', True)
         self.stacked = kwargs.pop('stacked', False)
-        self.ax_pos = np.arange(len(data)) + 0.25
-        if self.stacked:
-            self.tickoffset = 0.25
-        else:
-            self.tickoffset = 0.375
-        self.bar_width = 0.5
+
+        self.bar_width = kwargs.pop('width', 0.5)        
+        pos = kwargs.pop('position', 0.5)
+        self.ax_pos = np.arange(len(data)) + self.bar_width * pos
+
         self.log = kwargs.pop('log',False)
         MPLPlot.__init__(self, data, **kwargs)
 
+        if self.stacked or self.subplots:
+            self.tickoffset = self.bar_width * pos 
+        else:
+            K = self.nseries
+            self.tickoffset = self.bar_width * pos + self.bar_width / K
+
     def _args_adjust(self):
         if self.rot is None:
             self.rot = self._default_rot[self.kind]
@@ -1757,7 +1762,8 @@ def _make_plot(self):
                 pos_prior = pos_prior + np.where(mask, y, 0)
                 neg_prior = neg_prior + np.where(mask, 0, y)
             else:
-                rect = bar_f(ax, self.ax_pos + i * 0.75 / K, y, 0.75 / K,
+                w = self.bar_width / K
+                rect = bar_f(ax, self.ax_pos + (i + 1) * w, y, w,
                              start=start, label=label, **kwds)
             rects.append(rect)
             if self.mark_right:
@@ -1881,6 +1887,9 @@ def plot_frame(frame=None, x=None, y=None, subplots=False, sharex=True,
     colormap : str or matplotlib colormap object, default None
         Colormap to select colors from. If string, load colormap with that name
         from matplotlib.
+    position : float
+        Specify relative alignments for bar plot layout.
+        From 0 (left/bottom-end) to 1 (right/top-end). Default is 0.5 (center)
     kwds : keywords
         Options to pass to matplotlib plotting method
 
@@ -2016,6 +2025,9 @@ def plot_series(series, label=None, kind='line', use_index=True, rot=None,
     secondary_y : boolean or sequence of ints, default False
         If True then y-axis will be on the right
     figsize : a tuple (width, height) in inches
+    position : float
+        Specify relative alignments for bar plot layout.
+        From 0 (left/bottom-end) to 1 (right/top-end). Default is 0.5 (center)
     kwds : keywords
         Options to pass to matplotlib plotting method