BUG: Make xticks from _quarterly_finder() line up better (#47602)

* Make xticks from _quarterly_finder() line up better

When I plot a series of quarterly data that spans more than 11 years, my vertical grid lines get placed a year before where they should, i.e., one year before a year that is divisible by the default annual spacing. Changing one line in the _quarterly_finder() function, from +1 to +1970, fixes this for me. Can anyone please confirm the issue and that this fixes it?

* Create test for _quarterly_finder()

* Shorten comment

* Style refinements

* Style

* Sort imports

* Simplify test and move to test_converter.py file

* Newline at end of file

* Make 4 individual asserts

* Add Github discussion link as comment

* Remove newline from end of file

* Replace newline at end of file (minus tab)

* Add one line of white space

* Added release note in bug fixes/visualization section

* Moved release note to different section and noted public method

Author: elidourado

doc/source/whatsnew/v1.5.0.rst

@@ -976,6 +976,7 @@ Plotting
 - Bug in :meth:`DataFrame.plot.scatter` that prevented specifying ``norm`` (:issue:`45809`)
 - The function :meth:`DataFrame.plot.scatter` now accepts ``color`` as an alias for ``c`` and ``size`` as an alias for ``s`` for consistency to other plotting functions (:issue:`44670`)
 - Fix showing "None" as ylabel in :meth:`Series.plot` when not setting ylabel (:issue:`46129`)
+- Bug in :meth:`DataFrame.plot` that led to xticks and vertical grids being improperly placed when plotting a quarterly series (:issue:`47602`)
 
 Groupby/resample/rolling
 ^^^^^^^^^^^^^^^^^^^^^^^^

pandas/plotting/_matplotlib/converter.py

@@ -867,7 +867,8 @@ def _quarterly_finder(vmin, vmax, freq):
         info_fmt[year_start] = "%F"
 
     else:
-        years = dates_[year_start] // 4 + 1
+        # https://github.com/pandas-dev/pandas/pull/47602
+        years = dates_[year_start] // 4 + 1970
         nyears = span / periodsperyear
         (min_anndef, maj_anndef) = _get_default_annual_spacing(nyears)
         major_idx = year_start[(years % maj_anndef == 0)]

pandas/tests/plotting/test_converter.py

@@ -15,8 +15,10 @@
 from pandas import (
     Index,
     Period,
+    PeriodIndex,
     Series,
     Timestamp,
+    arrays,
     date_range,
 )
 import pandas._testing as tm
@@ -375,3 +377,32 @@ def get_view_interval(self):
         tdc = converter.TimeSeries_TimedeltaFormatter()
         monkeypatch.setattr(tdc, "axis", mock_axis())
         tdc(0.0, 0)
+
+
+@pytest.mark.parametrize("year_span", [11.25, 30, 80, 150, 400, 800, 1500, 2500, 3500])
+# The range is limited to 11.25 at the bottom by if statements in
+# the _quarterly_finder() function
+def test_quarterly_finder(year_span):
+    vmin = -1000
+    vmax = vmin + year_span * 4
+    span = vmax - vmin + 1
+    if span < 45:  # the quarterly finder is only invoked if the span is >= 45
+        return
+    nyears = span / 4
+    (min_anndef, maj_anndef) = converter._get_default_annual_spacing(nyears)
+    result = converter._quarterly_finder(vmin, vmax, "Q")
+    quarters = PeriodIndex(
+        arrays.PeriodArray(np.array([x[0] for x in result]), freq="Q")
+    )
+    majors = np.array([x[1] for x in result])
+    minors = np.array([x[2] for x in result])
+    major_quarters = quarters[majors]
+    minor_quarters = quarters[minors]
+    check_major_years = major_quarters.year % maj_anndef == 0
+    check_minor_years = minor_quarters.year % min_anndef == 0
+    check_major_quarters = major_quarters.quarter == 1
+    check_minor_quarters = minor_quarters.quarter == 1
+    assert np.all(check_major_years)
+    assert np.all(check_minor_years)
+    assert np.all(check_major_quarters)
+    assert np.all(check_minor_quarters)