BUG: Line plots aligned from start #57594 #57713

NoelMT · 2024-03-03T15:34:23Z

#57594 BUG: df.plot makes a shift to the right if frequency multiplies to n > 1

The cause of this bug is that by default all elements are generated to align to the end of
a period instead of at the start and the period technically does not end at the start of
the last period but rather just before the period subsequent to the last period. To fix this
problem we therefore passed a parameter stating that the elements should be aligned to
the start of the period instead of the end. Parameter "how=S" was added in pandas/plotting/_matplotlib/timeseries.py::maybe_convert_index(),
on the function call data.index.asfreq(freq=freq_str, how = "S").

Additional tests were added to test whether input data match data in plot and one test was merged into one the added test to reduce code duplication.

… the ones in the resulting plot * Added test_period_range_content which tests if plotted dates are equal to generated dates using period_range * Added test_date_range_content which tests if plotted dates are equal to generated dates using date_range

…time format

* Moved the checking of the plot data to a separate function * Added more frequencies to be tested

2-add-tests

…of the period

….py to test frequencies with koefficient > 1. Also removed redundant test called test_construction_from_string_monthly since the updated test_construction_from_string covers the same thing.

#2 Refactor: Updated test_construction_from_string::test_period_range…

datapythonista · 2024-03-05T19:04:55Z

Thank you for working on this @NoelMT. Do you mind adding plots to this PR (as a comment or in the description) that show how things are rendered before and after your changes please? That would give context to reviewers and it'll make reviewing much easier. Thanks!

natmokval · 2024-03-10T20:55:46Z

pandas/plotting/_matplotlib/timeseries.py

@@ -310,7 +310,7 @@ def maybe_convert_index(ax: Axes, data: NDFrameT) -> NDFrameT:
            if isinstance(data.index, ABCDatetimeIndex):
                data = data.tz_localize(None).to_period(freq=freq_str)
            elif isinstance(data.index, ABCPeriodIndex):
-                data.index = data.index.asfreq(freq=freq_str)
+                data.index = data.index.asfreq(freq=freq_str, how = "S") #Lineplot alignes to the start of period  


Could you please check ruff, ruff-format and codespell errors on line 313?
If you try the following command: pre-commit run --all-file, it should help fix these errors.

natmokval · 2024-03-10T20:57:59Z

pandas/tests/indexes/period/test_period_range.py

@@ -113,34 +119,6 @@ def test_construction_from_string(self, freq_offset, freq_period):
        result = period_range(start=end, end=start, freq=freq_period, name="foo")
        tm.assert_index_equal(result, expected)

-    def test_construction_from_string_monthly(self):


Thank you for noticing code duplication in pandas/tests/indexes/period/test_period_range.py. I am not sure how it’s related to this bug. Could you please move these changes to separate PR?

natmokval · 2024-03-10T21:01:46Z

pandas/tests/plotting/test_datetimelike.py

@@ -69,6 +70,31 @@ def test_fontsize_set_correctly(self):
        for label in ax.get_xticklabels() + ax.get_yticklabels():
            assert label.get_fontsize() == 2

+    @pytest.mark.parametrize("freq", ["20s","3s","60min","5min","7h","4D","8W","11M","2Q","Y", "3Y", "10Y"])


pre-commit run --all-file fails. Could you please check it?

natmokval · 2024-03-10T21:09:44Z

pandas/tests/plotting/test_datetimelike.py

@@ -69,6 +70,31 @@ def test_fontsize_set_correctly(self):
        for label in ax.get_xticklabels() + ax.get_yticklabels():
            assert label.get_fontsize() == 2

+    @pytest.mark.parametrize("freq", ["20s","3s","60min","5min","7h","4D","8W","11M","2Q","Y", "3Y", "10Y"])
+    def test_period_range_content(self, freq):
+        # Setup your DataFrame as in your scenario


instead of the comments, e.g. # Setup your DataFrame as in your scenario, could you please just add the number of GitHub issue: # GH#57594

natmokval · 2024-03-10T21:23:11Z

Thanks @NoelMT for working on this. Seems like it will fix the problem with visualisation of DataFrame if PeriodIndex as index provided.
e.g. for the first scenario:

from pandas import *
import numpy as np

idx = period_range("01/01/2000", freq='7h', periods=4)
df = DataFrame(
    np.array([0, 1, 0, 1]),
    index=idx,
    columns=["A"],
)
df.plot()
print(df)

plot before:

plot after:

natmokval · 2024-03-10T21:31:43Z

@NoelMT could you please add a note to doc/source/whatsnew/v3.0.0.rst in the section

pandas/doc/source/whatsnew/v3.0.0.rst

Line 356 in dc19148

Plotting

github-actions · 2024-04-10T00:05:25Z

This pull request is stale because it has been open for thirty days with no activity. Please update and respond to this comment if you're still interested in working on this.

mroeschke · 2024-04-15T17:24:18Z

Thanks for the pull request, but it appears to have gone stale. If interested in continuing, please merge in the main branch, address any review comments and/or failing tests, and we can reopen.

Gustaf-Larsson and others added 10 commits February 29, 2024 21:02

fix: 2-removed 100Y test since it is out of bounds timestamp for the …

5f8ad5e

…time format

feat: 2- Improved tests

f3ce3dc

* Moved the checking of the plot data to a separate function * Added more frequencies to be tested

Merge pull request #8 from NoelMT/2-AddTests

63b1f9f

2-add-tests

fix: #4 - Plotting PeriodIndexes now correctly aligns with the start …

d6eb450

…of the period

#2 Refactor: Updated test_construction_from_string::test_period_range…

e7e73e9

….py to test frequencies with koefficient > 1. Also removed redundant test called test_construction_from_string_monthly since the updated test_construction_from_string covers the same thing.

Merge pull request #9 from NoelMT/2-AddTests

19d794e

#2 Refactor: Updated test_construction_from_string::test_period_range…

Docs: added comments to bug fix #10

b3d4ff1

fix: Removed TODO comments

6e22a86

fix: Updated documentation

6a20930

NoelMT changed the title ~~BUG: Line plots alligned from start #57594~~ BUG: Line plots aligned from start #57594 Mar 4, 2024

datapythonista added the Visualization plotting label Mar 5, 2024

natmokval reviewed Mar 10, 2024

View reviewed changes

github-actions bot added the Stale label Apr 10, 2024

mroeschke closed this Apr 15, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

BUG: Line plots aligned from start #57594 #57713

BUG: Line plots aligned from start #57594 #57713

NoelMT commented Mar 3, 2024 •

edited

Loading

datapythonista commented Mar 5, 2024

natmokval Mar 10, 2024 •

edited

Loading

natmokval Mar 10, 2024 •

edited

Loading

natmokval Mar 10, 2024 •

edited

Loading

natmokval Mar 10, 2024

natmokval commented Mar 10, 2024

natmokval commented Mar 10, 2024

github-actions bot commented Apr 10, 2024

mroeschke commented Apr 15, 2024

BUG: Line plots aligned from start #57594 #57713

BUG: Line plots aligned from start #57594 #57713

Conversation

NoelMT commented Mar 3, 2024 • edited Loading

datapythonista commented Mar 5, 2024

natmokval Mar 10, 2024 • edited Loading

Choose a reason for hiding this comment

natmokval Mar 10, 2024 • edited Loading

Choose a reason for hiding this comment

natmokval Mar 10, 2024 • edited Loading

Choose a reason for hiding this comment

natmokval Mar 10, 2024

Choose a reason for hiding this comment

natmokval commented Mar 10, 2024

natmokval commented Mar 10, 2024

github-actions bot commented Apr 10, 2024

mroeschke commented Apr 15, 2024

NoelMT commented Mar 3, 2024 •

edited

Loading

natmokval Mar 10, 2024 •

edited

Loading

natmokval Mar 10, 2024 •

edited

Loading

natmokval Mar 10, 2024 •

edited

Loading