Inconsistent indexes for tick label plotting (#28733)

* TST: Test for issues #26186 and #11465

* BUG: Generate the tick position in BarPlot using convert tools from matlab.

Generate the tick position in BarPlot using convert tools from matlab.

* TST: Modify tests/plotting/test_frame.test_bar_categorical

Ticklocs are now float also for categorical bar data (as they are
position on the axis). The test is changed to compare to a array of
np.float.

* TST: Fix test for windows OS

* TST: Add test for plotting MultiIndex bar plot

A fix to issue #26186 revealed no tests existed about plotting a bar
plot for a MultiIndex, but a section of the user guide visualization
did. This section of the user guide is now in the test suite.

* BUG: Special case for MultiIndex bar plot

* DOC: Add whatsnew entry for PR #28733

* CLN: Clean up in code and doc

* CLN: Clean up test_bar_numeric

* DOC Move to whatsnew v1.1

* FIX: Make tick dtype int for backwards compatibility

* DOC: Improve whatsnew message

* ENH: Add UserWarning when plotting bar plot with MultiIndex

* CLN: Remove duplicate code line

* TST: Capture UserWarning for Bar plot with MultiIndex

* TST: Improve test explanation

* ENH: Raise UserWarning only if redrawing on existing axis with data

* DOC: Move to whatsnew v1.2.9

Co-authored-by: Marco Gorelli <[email protected]>

Author: nrebena

doc/source/whatsnew/v1.2.0.rst

@@ -668,6 +668,7 @@ Plotting
 
 - Bug in :meth:`DataFrame.plot` was rotating xticklabels when ``subplots=True``, even if the x-axis wasn't an irregular time series (:issue:`29460`)
 - Bug in :meth:`DataFrame.plot` where a marker letter in the ``style`` keyword sometimes causes a ``ValueError`` (:issue:`21003`)
+- Bug in :func:`DataFrame.plot.bar` and :func:`Series.plot.bar`. Ticks position were assigned by value order instead of using the actual value for numeric, or a smart ordering for string. (:issue:`26186` and :issue:`11465`)
 - Twinned axes were losing their tick labels which should only happen to all but the last row or column of 'externally' shared axes (:issue:`33819`)
 - Bug in :meth:`Series.plot` and :meth:`DataFrame.plot` was throwing :exc:`ValueError` with a :class:`Series` or :class:`DataFrame`
   indexed by a :class:`TimedeltaIndex` with a fixed frequency when x-axis lower limit was greater than upper limit (:issue:`37454`)

pandas/plotting/_matplotlib/core.py

@@ -1359,7 +1359,6 @@ def __init__(self, data, **kwargs):
         self.bar_width = kwargs.pop("width", 0.5)
         pos = kwargs.pop("position", 0.5)
         kwargs.setdefault("align", "center")
-        self.tick_pos = np.arange(len(data))
 
         self.bottom = kwargs.pop("bottom", 0)
         self.left = kwargs.pop("left", 0)
@@ -1382,7 +1381,16 @@ def __init__(self, data, **kwargs):
                 self.tickoffset = self.bar_width * pos
                 self.lim_offset = 0
 
-        self.ax_pos = self.tick_pos - self.tickoffset
+        if isinstance(self.data.index, ABCMultiIndex):
+            if kwargs["ax"] is not None and kwargs["ax"].has_data():
+                warnings.warn(
+                    "Redrawing a bar plot with a MultiIndex is not supported "
+                    + "and may lead to inconsistent label positions.",
+                    UserWarning,
+                )
+            self.ax_index = np.arange(len(data))
+        else:
+            self.ax_index = self.data.index
 
     def _args_adjust(self):
         if is_list_like(self.bottom):
@@ -1409,6 +1417,15 @@ def _make_plot(self):
 
         for i, (label, y) in enumerate(self._iter_data(fillna=0)):
             ax = self._get_ax(i)
+
+            if self.orientation == "vertical":
+                ax.xaxis.update_units(self.ax_index)
+                self.tick_pos = ax.convert_xunits(self.ax_index).astype(np.int)
+            elif self.orientation == "horizontal":
+                ax.yaxis.update_units(self.ax_index)
+                self.tick_pos = ax.convert_yunits(self.ax_index).astype(np.int)
+            self.ax_pos = self.tick_pos - self.tickoffset
+
             kwds = self.kwds.copy()
             if self._is_series:
                 kwds["color"] = colors
@@ -1480,8 +1497,8 @@ def _post_plot_logic(self, ax: "Axes", data):
             str_index = [pprint_thing(key) for key in range(data.shape[0])]
         name = self._get_index_name()
 
-        s_edge = self.ax_pos[0] - 0.25 + self.lim_offset
-        e_edge = self.ax_pos[-1] + 0.25 + self.bar_width + self.lim_offset
+        s_edge = self.ax_pos.min() - 0.25 + self.lim_offset
+        e_edge = self.ax_pos.max() + 0.25 + self.bar_width + self.lim_offset
 
         self._decorate_ticks(ax, name, str_index, s_edge, e_edge)
 

pandas/tests/plotting/frame/test_frame.py

@@ -2190,6 +2190,81 @@ def test_xlabel_ylabel_dataframe_plane_plot(self, kind, xlabel, ylabel):
         assert ax.get_xlabel() == (xcol if xlabel is None else xlabel)
         assert ax.get_ylabel() == (ycol if ylabel is None else ylabel)
 
+    @pytest.mark.slow
+    @pytest.mark.parametrize("method", ["bar", "barh"])
+    def test_bar_ticklabel_consistence(self, method):
+        # Draw two consecutiv bar plot with consistent ticklabels
+        # The labels positions should not move between two drawing on the same axis
+        # GH: 26186
+        def get_main_axis(ax):
+            if method == "barh":
+                return ax.yaxis
+            elif method == "bar":
+                return ax.xaxis
+
+        # Plot the first bar plot
+        data = {"A": 0, "B": 3, "C": -4}
+        df = DataFrame.from_dict(data, orient="index", columns=["Value"])
+        ax = getattr(df.plot, method)()
+        ax.get_figure().canvas.draw()
+
+        # Retrieve the label positions for the first drawing
+        xticklabels = [t.get_text() for t in get_main_axis(ax).get_ticklabels()]
+        label_positions_1 = dict(zip(xticklabels, get_main_axis(ax).get_ticklocs()))
+
+        # Modify the dataframe order and values and plot on same axis
+        df = df.sort_values("Value") * -2
+        ax = getattr(df.plot, method)(ax=ax, color="red")
+        ax.get_figure().canvas.draw()
+
+        # Retrieve the label positions for the second drawing
+        xticklabels = [t.get_text() for t in get_main_axis(ax).get_ticklabels()]
+        label_positions_2 = dict(zip(xticklabels, get_main_axis(ax).get_ticklocs()))
+
+        # Assert that the label positions did not change between the plotting
+        assert label_positions_1 == label_positions_2
+
+    def test_bar_numeric(self):
+        # Bar plot with numeric index have tick location values equal to index
+        # values
+        # GH: 11465
+        df = DataFrame(np.random.rand(10), index=np.arange(10, 20))
+        ax = df.plot.bar()
+        ticklocs = ax.xaxis.get_ticklocs()
+        expected = np.arange(10, 20, dtype=np.int64)
+        tm.assert_numpy_array_equal(ticklocs, expected)
+
+    def test_bar_multiindex(self):
+        # Test from pandas/doc/source/user_guide/visualization.rst
+        # at section Plotting With Error Bars
+        # Related to issue GH: 26186
+
+        ix3 = pd.MultiIndex.from_arrays(
+            [
+                ["a", "a", "a", "a", "b", "b", "b", "b"],
+                ["foo", "foo", "bar", "bar", "foo", "foo", "bar", "bar"],
+            ],
+            names=["letter", "word"],
+        )
+
+        df3 = DataFrame(
+            {"data1": [3, 2, 4, 3, 2, 4, 3, 2], "data2": [6, 5, 7, 5, 4, 5, 6, 5]},
+            index=ix3,
+        )
+
+        # Group by index labels and take the means and standard deviations
+        # for each group
+        gp3 = df3.groupby(level=("letter", "word"))
+        means = gp3.mean()
+        errors = gp3.std()
+
+        # No assertion we just ensure that we can plot a MultiIndex bar plot
+        # and are getting a UserWarning if redrawing
+        with tm.assert_produces_warning(None):
+            ax = means.plot.bar(yerr=errors, capsize=4)
+        with tm.assert_produces_warning(UserWarning):
+            means.plot.bar(yerr=errors, capsize=4, ax=ax)
+
 
 def _generate_4_axes_via_gridspec():
     import matplotlib as mpl