ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part III) #19884

topper-123 · 2018-02-24T17:10:42Z

closes ENH: Let Series/DataFrame initialisation from dicts use insertion order when python>=3.6 #19018
xref ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part I) #19830 and ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part II) #19859
tests added / passed
passes git diff upstream/master -u -- "*.py" | flake8 --diff

This PR lets pandas allow initialisation using insertion order for python>=3.6. E.g. Series({'B': 1, 'A': 2}) will retain order and not convert to Series({'A': 2, 'B': 1}). As such this changes the API soemwhat fundamentally (but in a very good way, IMO).

Some questions:

Should this change be at the top of the whatsnew, in order to increase visibility to this change?
Should an option be added to pd.options, so user can revert back to the old behaviour? I'm thinking here of users who have made tests based on dict sort-by-values, and whose tests will break by this new change
Should there be a deprecation procedure, as asked by @gfyoung ?

I also need to look thorugh the documentation if changes should be made there, but would appreciate getting the basic PR approved first.

jreback · 2018-02-24T17:12:40Z

doc/source/whatsnew/v0.23.0.txt

+
+.. code-block:: ipython
+
+  In [1]: pd.Series({'Income': 2000,


you need another space here

jreback · 2018-02-24T17:13:37Z

pandas/tests/frame/test_constructors.py

@@ -241,6 +241,17 @@ def test_constructor_dict(self):
        # Corner cases
        assert len(DataFrame({})) == 0



make new tests

Can you be more specific?

make these as new tests, don't add onto the existing ones (IOW give it a separate test name)

jreback · 2018-02-24T17:13:42Z

pandas/tests/series/test_constructors.py

@@ -775,6 +775,17 @@ def test_constructor_dict(self):
        expected = Series([1, 2, nan, 0], index=['b', 'c', 'd', 'a'])
        assert_series_equal(result, expected)



jreback · 2018-02-24T17:14:03Z

pandas/tests/series/test_constructors.py

+        # GH19018
+        # initialization ordering: by insertion order if python>= 3.6, else
+        # by value
+        d = {'b': 1, 'a': 0, 'c': 2}


can you add ones for SparseSeries/DataFrame

topper-123 · 2018-02-24T19:27:26Z

Comments addressed.

jschendel · 2018-02-24T20:49:17Z

doc/source/whatsnew/v0.23.0.txt

@@ -240,7 +240,7 @@ The :func:`DataFrame.assign` now accepts dependent keyword arguments for python
  using ``.assign()`` to update an existing column. Previously, callables
  referring to other variables being updated would get the "old" values

-  Previous Behaviour:
+  Previous behaviour:


It looks like we generally default to American English spellings within the codebase, so could do "behavior" here instead. Found many instances of "behaviour" in the codebase though, but still far outnumbered by "behavior".

jschendel · 2018-02-24T20:49:23Z

doc/source/whatsnew/v0.23.0.txt

@@ -253,7 +253,7 @@ The :func:`DataFrame.assign` now accepts dependent keyword arguments for python
      1  3 -2
      2  4 -3

-  New Behaviour:
+  New behaviour:


jschendel · 2018-02-24T20:49:32Z

doc/source/whatsnew/v0.23.0.txt

+Pandas will from version 0.23 use insertion order, when creating series or
+data frames from dicts (:issue:`19018`) .
+
+Previous behaviour (and current behaviour if on Python < 3.6):


jschendel · 2018-02-24T20:49:37Z

doc/source/whatsnew/v0.23.0.txt

+
+Note the series above is ordered alphabetically by the index values.
+
+New behaviour (for Python >= 3.6):


jschendel · 2018-02-24T20:49:58Z

doc/source/whatsnew/v0.23.0.txt

+               'Taxes': -200,
+               'Net result': 300})
+
+Notice that the series is now ordered by insertion order. This new behaviour is


jschendel · 2018-02-24T20:50:04Z

doc/source/whatsnew/v0.23.0.txt

+used for all relevant pandas types (``Series``, ``DataFrame``, ``SparseSeries``
+and ``SparseDataFrame``).
+
+If you wish to retain the old behaviour while using Python >= 3.6, you can use


jschendel · 2018-02-24T20:50:54Z

doc/source/whatsnew/v0.23.0.txt

@@ -320,6 +320,57 @@ If installed, we now require:
 | openpyxl        | 2.4.0           |          |
 +-----------------+-----------------+----------+

+.. _whatsnew_0230.api_breaking.dict_insertion_order:
+
+Creating dataframes and series from dicts preserves dict insertion order for python 3.6+


Some capitalization: dataframes --> DataFrames, series --> Series, python --> Python

jschendel · 2018-02-24T20:56:19Z

pandas/tests/series/test_constructors.py

@@ -783,6 +783,18 @@ def test_constructor_dict(self):
        expected.iloc[1] = 1
        assert_series_equal(result, expected)

+    def test_constructor_dict_order(self):


Maybe could decorate tests like this one with @pytest.mark.skipif(not PY36, reason="Python 3.6 specific behavior") and only test the PY36 path. Not sure if we're gaining much testing the non-PY36 path. Would wait to see what others think before making this change though.

Pandas has clearly defined ordering for python < 3.6 also (sort by values). So it's nneeded to test this distinction and not use a skipif, IMO.

@topper-123 let's go ahead and make 2 different tests here, 1 for >= 3.6 and one for less, they should skip on the alternative, same for the tests in frame.

jschendel · 2018-02-24T21:15:18Z

doc/source/whatsnew/v0.23.0.txt

+version 3.6 and later have changed the ordering definition of dicts, so dicts
+in these newer versions are ordered by insertion order
+(see also `PEP 468 <https://www.python.org/dev/peps/pep-0468/>`_).
+Pandas will from version 0.23 use insertion order, when creating series or


series --> Series

jschendel · 2018-02-24T21:15:30Z

doc/source/whatsnew/v0.23.0.txt

+in these newer versions are ordered by insertion order
+(see also `PEP 468 <https://www.python.org/dev/peps/pep-0468/>`_).
+Pandas will from version 0.23 use insertion order, when creating series or
+data frames from dicts (:issue:`19018`) .


data frames --> DataFrames

topper-123 · 2018-02-25T09:26:32Z

I've uploaded changes, incl. documentation changes.

codecov · 2018-02-25T13:34:37Z

Codecov Report

Merging #19884 into master will decrease coverage by 0.01%.
The diff coverage is 91.66%.

@@            Coverage Diff             @@
##           master   #19884      +/-   ##
==========================================
- Coverage   91.69%   91.67%   -0.02%     
==========================================
  Files         150      150              
  Lines       49043    49045       +2     
==========================================
- Hits        44968    44963       -5     
- Misses       4075     4082       +7

Flag	Coverage Δ
#multiple	`90.05% <91.66%> (-0.02%)`	⬇️
#single	`41.79% <91.66%> (-0.04%)`	⬇️

Impacted Files	Coverage Δ
pandas/core/sparse/series.py	`95.2% <ø> (ø)`	⬆️
pandas/core/frame.py	`97.18% <100%> (-0.01%)`	⬇️
pandas/core/series.py	`93.85% <100%> (-0.39%)`	⬇️
pandas/core/sparse/frame.py	`94.82% <100%> (+0.01%)`	⬆️
pandas/core/panel.py	`97.29% <100%> (-0.01%)`	⬇️
pandas/core/common.py	`92.04% <83.33%> (-0.92%)`	⬇️

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d615f86...473f564. Read the comment docs.

jreback · 2018-02-25T15:49:10Z

doc/source/dsintro.rst

-If ``data`` is a dict, if **index** is passed the values in data corresponding
-to the labels in the index will be pulled out. Otherwise, an index will be
-constructed from the sorted keys of the dict, if possible.
+When creating a pandas Series from a dict, the Series will be ordered by the


add a versionadded tag

jreback · 2018-02-25T15:49:36Z

doc/source/dsintro.rst

@@ -277,6 +289,8 @@ The row and column labels can be accessed respectively by accessing the
 From dict of ndarrays / lists
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

+The columns will be ordered by the dict insertion order, unless you're using


add this in a ::note..

jreback · 2018-02-25T15:49:54Z

doc/source/dsintro.rst

@@ -277,6 +289,8 @@ The row and column labels can be accessed respectively by accessing the
 From dict of ndarrays / lists
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

+The columns will be ordered by the dict insertion order, unless you're using
+Python version < 3.6, then the columns will be ordered lexically/alphabetically.


just specify lexically

jreback · 2018-02-25T15:50:34Z

doc/source/whatsnew/v0.23.0.txt

@@ -240,7 +240,7 @@ The :func:`DataFrame.assign` now accepts dependent keyword arguments for python
  using ``.assign()`` to update an existing column. Previously, callables
  referring to other variables being updated would get the "old" values

-  Previous Behaviour:
+  Previous behavior:


hmm, is the capitalization consistent across all of these entries? rather you leave this alone for now (and we can come back and clean it up later)

jreback · 2018-02-25T15:50:47Z

doc/source/whatsnew/v0.23.0.txt

@@ -320,6 +320,57 @@ If installed, we now require:
 | openpyxl        | 2.4.0           |          |
 +-----------------+-----------------+----------+

+.. _whatsnew_0230.api_breaking.dict_insertion_order:
+
+Creating Dataframes and Series from dicts preserves dict insertion order for python 3.6+


make this shorter title

jreback · 2018-02-25T15:51:31Z

doc/source/whatsnew/v0.23.0.txt

+Until Python 3.6, dicts in Python had no formally defined ordering. Python
+version 3.6 and later have changed the ordering definition of dicts, so dicts
+in these newer versions are ordered by insertion order
+(see also `PEP 468 <https://www.python.org/dev/peps/pep-0468/>`_).


don't use parens in the 'see also', rather highlite some relevant phrase with this reference.

jreback · 2018-02-25T15:51:43Z

doc/source/whatsnew/v0.23.0.txt

+version 3.6 and later have changed the ordering definition of dicts, so dicts
+in these newer versions are ordered by insertion order
+(see also `PEP 468 <https://www.python.org/dev/peps/pep-0468/>`_).
+Pandas will from version 0.23 use insertion order, when creating Series or


you don't need from version 0.23, this is the whatsnew

jreback · 2018-02-25T15:52:13Z

doc/source/whatsnew/v0.23.0.txt

+.. code-block:: ipython
+
+   In [1]: pd.Series({'Income': 2000,
+   ...                 'Expenses': -1500,


this is not aligned

jreback · 2018-02-25T15:54:53Z

pandas/tests/series/test_constructors.py

@@ -783,6 +783,18 @@ def test_constructor_dict(self):
        expected.iloc[1] = 1
        assert_series_equal(result, expected)

+    def test_constructor_dict_order(self):


@topper-123 let's go ahead and make 2 different tests here, 1 for >= 3.6 and one for less, they should skip on the alternative, same for the tests in frame.

jreback · 2018-02-25T15:55:49Z

pandas/core/sparse/frame.py

@@ -138,7 +142,10 @@ def _init_dict(self, data, index, columns, dtype=None):
            columns = _ensure_index(columns)
            data = {k: v for k, v in compat.iteritems(data) if k in columns}
        else:
-            columns = Index(com._try_sort(list(data.keys())))
+            keys = list(data.keys())


try pushing this entirely to _try_sort, see if anything breaks as this used in a couple of places and will provide some 3.6 consistency.

_try_sort is used for sorting lists in a couple of places.

I don't think it's a good idea to have a sort function that doesn't sort on ordered dict keys, but sorts on ordered lists.

Maybe a differently named function, e.g. _dict_keys_to_ordered_list?

ok with renaming this, but do try to change all of the existing functions to use it.
You are creating an inconsistency by making dict inputs follow insertion order, rather have all instances fixed.

I've changed this for all places where _try_sort is used to sort list of dict keys. New function _dict_keys_to_ordered_list added.

pep8speaks · 2018-02-28T17:46:02Z

Hello @topper-123! Thanks for updating the PR.

Cheers ! There are no PEP8 issues in this Pull Request. 🍻

Comment last updated on March 02, 2018 at 11:18 Hours UTC

TomAugspurger · 2018-02-28T20:14:37Z

doc/source/dsintro.rst

+   d = {'b' : 1, 'a' : 0, 'c' : 2}
+   pd.Series(d)
+
+If you in the example above were on a Python version lower than 3.6 or Pandas


typo: I think "you in" can be removed

TomAugspurger · 2018-02-28T20:18:58Z

doc/source/whatsnew/v0.23.0.txt

+
+Until Python 3.6, dicts in Python had no formally defined ordering. Python
+version 3.6 and later have changed the ordering definition of dicts, so dicts
+in these newer versions are ordered by insertion order, see


Change the second sentence to "For Python 3.6 and later, dictionaries are ordered by insertion order."?

topper-123 · 2018-03-01T08:59:35Z

The errors in travis-ci were

1:
One which passed locally, so sure how to handle that. It complains about resources also, so maybe the failures are related to that?

2:
One with the exit quote:

The job exceeded the maximum time limit for jobs, and has been terminated.

so unrelated to my PR.

Maybe the tests should be run once again?

topper-123 · 2018-03-01T20:02:44Z

Green.

jorisvandenbossche

small doc comment

jorisvandenbossche · 2018-03-01T20:41:34Z

doc/source/dsintro.rst

-If ``data`` is a dict, if **index** is passed the values in data corresponding
-to the labels in the index will be pulled out. Otherwise, an index will be
-constructed from the sorted keys of the dict, if possible.
+.. note::


Can you move this note a bit below? It seems strange to start this "from dict" section with a note without first explaining what it actually is.
So I would give an example first, and only afterwards noting that the output depends on the python version

jorisvandenbossche · 2018-03-01T20:46:21Z

doc/source/dsintro.rst

@@ -243,12 +260,21 @@ not matching up to the passed index.
 If axis labels are not passed, they will be constructed from the input data
 based on common sense rules.

+.. note::
+
+   When the data is a dict, and columns is not passed, the DataFrame columns


"and columns is not passed" -> I would try to make it clear that it is about a keyword columns, by either qouting it, or "no columns are passed", or "columns is not specified"

topper-123 · 2018-03-02T09:32:37Z

Comments from @jorisvandenbossche addressed.

jorisvandenbossche · 2018-03-02T10:49:24Z

@topper-123 For future reference, could you add commits instead of amending it to the previous or squashing? That makes it easier to see what you updated.

jorisvandenbossche

Changes look good!

jreback · 2018-03-02T11:19:11Z

thanks @topper-123

@TomAugspurger

commit df2e361 Author: Jeff Reback <jeff@reback.net> Date: Sun Mar 11 18:33:25 2018 -0400 LINT: fixing commit f1c0b7c Author: David Polo <delkk0@users.noreply.github.com> Date: Sun Mar 11 22:54:27 2018 +0100 DOC: Improved the docstring of pandas.plotting._core.FramePlotMethods… (pandas-dev#20157) * DOC: Improved the docstring of pandas.plotting._core.FramePlotMethods.barh() - Added examples section - Added extended summary - Added argument explanation * DOC: Improved the docstring of pandas.plotting._core.FramePlotMethods.barh() - Correcting PR comments * DOC: Improved the docstring of pandas.plotting._core.FramePlotMethods.barh() - Adding defaults for variables. * Update reference commit 0780193 Author: Jonas Schulze <jonas.schulze7@t-online.de> Date: Sun Mar 11 22:37:37 2018 +0100 DOC: update the pandas.DataFrame.plot.density docstring (pandas-dev#20236) * DOC: update the pandas.DataFrame.plot.kde and pandas.Series.plot.kde docstrings Unfortunately, I was not able to compute a kernel estimate of a two-dimensional random variable. Hence, the example is more of an analysis of some independent data series. * DOC: extract similarities of kde docstrings The `DataFrame.plot.kde` and `Series.plot.kde` now use a common docstring, for which the differences are inserted. commit 2718984 Author: Cihan Ceyhan <chncyhn@gmail.com> Date: Sun Mar 11 21:48:08 2018 +0100 DOC: Update the pandas.Series.dt.round/floor/ceil docstrings (pandas-dev#20187) * DOC: Update the pandas.Series.dt.round/floor/ceil docstrings * DOC: review points fixed. * Add series commit 0d86742 Author: Antonio Molina <aydevosotros@gmail.com> Date: Sun Mar 11 18:57:37 2018 +0100 DOC: Improved pandas.plotting.bootstrap_plot docstring (pandas-dev#20166) * Improved documentation on bootstrap_plot * Improved documentation on bootstrap_plot * Doc bootstrap_plot: Fixed some comments on pull requests * Added reference to wikipedia * Changed kwds for **kwds * Removed ** from kwds becuase of validation iuses * Fixed forgotten break line. I think that the kwds paramater now fits what expected @TomAugspurger. If not, sorry and indicate how it should be * Fixed warnings on compilation * Moved reference to extended description commit a2910ad Author: András Novoszáth <nocibambi@gmail.com> Date: Sun Mar 11 18:56:01 2018 +0100 DOC: update the Index.get_values docstring (pandas-dev#20231) * DOC: update the Index.get_values docstring * Corrections * Corrected extended summary and quotes * Correcting spaces, extended summary, multiIndex example * See also correction * Multi ndim commit afa6c42 Author: Marc <mlafore05@gmail.com> Date: Sun Mar 11 10:42:35 2018 -0400 DOC: update the pandas.DataFrame.all docstring (pandas-dev#20216) commit a44bae3 Author: Victor Villas <villasv@outlook.com> Date: Sun Mar 11 11:41:12 2018 -0300 DOC: update the Series.view docstring (pandas-dev#20220) commit 233103f Author: David Adrián Cañones Castellano <davidarcano@gmail.com> Date: Sun Mar 11 15:40:02 2018 +0100 DOC: update the docstring of pandas.DataFrame.from_dict (pandas-dev#20259) commit 62bddec Author: csfarkas <csaba.farkas95@gmail.com> Date: Sun Mar 11 15:33:54 2018 +0100 DOC: add docstring for Index.get_duplicates (pandas-dev#20223) commit 8c77238 Author: adatasetaday <32177771+adatasetaday@users.noreply.github.com> Date: Sun Mar 11 10:17:05 2018 -0400 Docstring pandas.series.diff (pandas-dev#20238) commit 4271757 Author: Aly Sivji <4369343+alysivji@users.noreply.github.com> Date: Sun Mar 11 08:51:25 2018 -0500 DOC: update `pandas/core/ops.py` docstring template to accept examples (pandas-dev#20246) commit 080ef0c Author: akosel <aaronjkosel@gmail.com> Date: Sun Mar 11 12:43:10 2018 +0000 DOC: update the DataFrame.iat[] docstring (pandas-dev#20219) * DOC: update the DataFrame.iat[] docstring * Update based on PR comments * Update based on PR comments * Singular not plural * Update to account for use with Series. Add example using Series. * Update indexing.py * PEP8 commit 302fda4 Author: adatasetaday <32177771+adatasetaday@users.noreply.github.com> Date: Sun Mar 11 08:36:21 2018 -0400 DOC: update the pandas.DataFrame.diff docstring (pandas-dev#20227) * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * DOC: update the pandas.DataFrame.diff docstring * Cleanup commit c791a84 Author: Pietro Battiston <me@pietrobattiston.it> Date: Sun Mar 11 13:07:01 2018 +0100 DOC: pd.core.window.Expanding.kurt docstring (split from pd.core.Rolling.kurt) (pandas-dev#20064) commit b3d6ce6 Author: Nipun Sadvilkar <nipunsadvilkar@gmail.com> Date: Sun Mar 11 17:29:33 2018 +0530 DOC: update the pandas.date_range() docstring (pandas-dev#20143) * DOC: Improved the docstring of pandas.date_range() * Change date strings to iso format * Removed import pands in Examples docstring * Add See Also Docstring * Update datetimes.py * Doctests commit 6d7272a Author: Samuel Sinayoko <samuelsinayoko@bmlltech.com> Date: Sun Mar 11 11:58:09 2018 +0000 DOC: update DataFrame.to_records (pandas-dev#20191) * Update to_records docstring. - Minor changes (missing dots, newlines) to make tests pass. - More examples. * Fix html docs. Missing newlines. * Reword datetime type information. * flake8 errors * Fix typo (duplicated type) * Remove unwanted blank line after Examples. * Fix doctests. ``` (pandas_dev) sinayoks@landade:~/dev/pandas/ $ pytest --doctest-modules pandas/core/frame.py -k to_record ========================================================================================== test session starts ========================================================================================== platform darwin -- Python 3.6.4, pytest-3.4.2, py-1.5.2, pluggy-0.6.0 rootdir: /Users/sinayoks/dev/pandas, inifile: setup.cfg plugins: xdist-1.22.1, forked-0.2, cov-2.5.1 collected 43 items pandas/core/frame.py . [100%] ========================================================================================== 42 tests deselected ========================================================================================== ``` * Few more changes commit 636335a Author: Gabriel de Maeztu <gabriel.maeztu@gmail.com> Date: Sun Mar 11 12:56:48 2018 +0100 DOC: Improved the docstring of pandas.plotting.radviz (pandas-dev#20169) commit fbebc7f Author: jen w <j.e.weiss@gmail.com> Date: Sun Mar 11 06:50:54 2018 -0500 DOC: Update pandas.DataFrame.tail docstring (pandas-dev#20225) commit c2864d7 Author: Stephen Childs <sechilds@gmail.com> Date: Sun Mar 11 07:50:39 2018 -0400 DOC: update the DataFrame.cov docstring (pandas-dev#20245) * DOC: Revise docstring of DataFrame cov method Update the docstring with some examples from elsewhere in the pandas documentation. Some of the examples use randomly generated time series because we need to get covariance between long series. Used a random seed to ensure that the results are the same each time. * DOC: Fix See Also and min_periods explanation. Responding to comments on PR. See also section will link properly and number of periods explanation clearer. commit 90e31b9 Author: jen w <j.e.weiss@gmail.com> Date: Sun Mar 11 06:50:18 2018 -0500 DOC: update pandas.DataFrame.head docstring (pandas-dev#20262) commit fb556ed Author: Israel Saeta Pérez <dukebody@gmail.com> Date: Sat Mar 10 22:33:42 2018 +0100 DOC: Improve pandas.Series.plot.kde docstring and kwargs rewording for whole file (pandas-dev#20041) commit c3d491a Author: Andy R. Terrel <andy.terrel@gmail.com> Date: Sat Mar 10 11:48:13 2018 -0800 DOC: update the DataFrame.head() docstring (pandas-dev#20206) commit dd7f567 Author: DataOmbudsman <DataOmbudsman@users.noreply.github.com> Date: Sat Mar 10 20:15:48 2018 +0100 DOC: update the Index.shift docstring (pandas-dev#20192) * DOC: updating docstring of Index.shift * Add See Also section to shift * Update link to Series.shift commit 5b0caf4 Author: Eric O. LEBIGOT (EOL) <lebigot@users.noreply.github.com> Date: Sat Mar 10 17:32:20 2018 +0100 DOC: update the Series.memory_usage() docstring (pandas-dev#20086) commit 9fb7ac9 Author: Carol Willing <carolcode@willingconsulting.com> Date: Sat Mar 10 08:28:54 2018 -0800 DOC: Edit contributing to docs section (pandas-dev#20190) commit d8181a5 Author: DaanVanHauwermeiren <DaanVanHauwermeiren@users.noreply.github.com> Date: Sat Mar 10 17:25:20 2018 +0100 DOC: update the Series.isin docstring (pandas-dev#20175) commit ec631ce Author: Riccardo Magliocchetti <riccardo.magliocchetti@gmail.com> Date: Sat Mar 10 17:12:41 2018 +0100 DOC: update the pandas.Series.tail docstring (pandas-dev#20176) commit e5e4ae9 Author: DaanVanHauwermeiren <DaanVanHauwermeiren@users.noreply.github.com> Date: Sat Mar 10 16:41:58 2018 +0100 DOC: update the pandas.Index.drop_duplicates and pandas.Series.drop_duplicates docstring (pandas-dev#20114) commit d7bcb22 Author: Riccardo Magliocchetti <riccardo.magliocchetti@gmail.com> Date: Sat Mar 10 15:49:31 2018 +0100 DOC: update the MultiIndex.swaplevel docstring (pandas-dev#20105) commit 8497029 Author: Gjelt <math-and-data@users.noreply.github.com> Date: Sat Mar 10 15:41:17 2018 +0100 DOC: Improved the docstring of pandas.DataFrame.values (pandas-dev#20065) commit 840d432 Author: Jordi Contestí <25779507+jcontesti@users.noreply.github.com> Date: Sat Mar 10 13:24:35 2018 +0100 DOC: Improved the docstring of Series.str.findall (pandas-dev#19982) commit 2a0d23b Author: Jeff Reback <jeff@reback.net> Date: Sat Mar 10 06:54:19 2018 -0500 DOC: lint commit bf0dcb5 Author: Kate Surta <kate.surta@gmail.com> Date: Sat Mar 10 14:42:52 2018 +0300 BUG: Check for wrong arguments in index subclasses constructors (pandas-dev#20017) commit 4131149 Author: Stijn Van Hoey <stijnvanhoey@gmail.com> Date: Sat Mar 10 10:15:41 2018 +0100 DOC: Extend docstring pandas core index to_frame method (pandas-dev#20036) commit 52cffa3 Author: William Ayd <william.ayd@icloud.com> Date: Fri Mar 9 18:06:43 2018 -0800 Cythonized GroupBy pct_change (pandas-dev#19919) commit da6f827 Author: William Ayd <william.ayd@icloud.com> Date: Fri Mar 9 18:03:50 2018 -0800 Refactored GroupBy ASVs (pandas-dev#20043) commit bd31f71 Author: William Ayd <william.ayd@icloud.com> Date: Fri Mar 9 17:53:34 2018 -0800 Added 'displayed_only' option to 'read_html' (pandas-dev#20047) commit ed96567 Author: Ksenia <bobrovaksenia@gmail.com> Date: Sat Mar 10 02:40:10 2018 +0100 TST: series/indexing tests parametrization + moving test methods (pandas-dev#20059) commit 7c14e4f Author: Kyle Barron <kylebarron2@gmail.com> Date: Fri Mar 9 11:31:14 2018 -0500 DOC: Add syntax highlighting to SAS code blocks in comparison_with_sas.rst (pandas-dev#20080) * Add syntax highlighting to SAS code blocks * Fix typo commit 731d971 Author: Matthew Roeschke <emailformattr@gmail.com> Date: Fri Mar 9 03:30:22 2018 -0800 Fix typo in apply.py (pandas-dev#20058) commit cc1b934 Author: Matthew Roeschke <emailformattr@gmail.com> Date: Fri Mar 9 03:13:50 2018 -0800 BUG: Retain timezone dtype with cut and qcut (pandas-dev#19890) commit c730d08 Author: William Ayd <william.ayd@icloud.com> Date: Fri Mar 9 02:37:27 2018 -0800 DOC: Update Kurt Docstr (pandas-dev#20044) commit 9119d07 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Fri Mar 9 10:03:44 2018 +0100 Temporary github PR template for sprint (pandas-dev#20055) commit 747501a Author: Aly Sivji <4369343+alysivji@users.noreply.github.com> Date: Fri Mar 9 02:19:59 2018 -0600 DOC: Improve docstring for pandas.Index.repeat (pandas-dev#19985) commit 1d73cf3 Author: Rouz Azari <rouzazari@users.noreply.github.com> Date: Thu Mar 8 16:54:53 2018 -0800 BUG: Dense ranking with percent now uses 100% basis (pandas-dev#15639) commit f9fd540 Author: William Ayd <william.ayd@icloud.com> Date: Thu Mar 8 16:36:23 2018 -0800 Added flake8 to DEV requirements (pandas-dev#20063) commit b669112 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Thu Mar 8 14:09:12 2018 +0100 DOC: require returns section in validation script (pandas-dev#19994) commit 024d8b4 Author: Jeff Reback <jeff@reback.net> Date: Thu Mar 8 07:08:57 2018 -0500 TST: xfail test_time on py2 & mpl 1.4.3 (pandas-dev#20053) commit b85f6c1 Author: Marc Garcia <garcia.marc@gmail.com> Date: Thu Mar 8 11:07:08 2018 +0000 DOC: update docstring validation script + replace api coverage script (pandas-dev#20025) * Improvments to validate_docstrings script: adding sections to summary, validating type and description of parameters * DOC: Improvements to validate docstring script (added api_coverage functionality, sections in csv and extra validations) commit 9273bf5 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Thu Mar 8 11:14:05 2018 +0100 DOC/CI: temp pin matplotlib for doc build (pandas-dev#20045) commit 63ce781 Author: Jeff Reback <jeff@reback.net> Date: Wed Mar 7 17:01:38 2018 -0500 TST: xfail mpl 2.2 tests xref pandas-dev#20031 commit 7c7bd56 Author: Daniel Frank <danfrankj@gmail.com> Date: Wed Mar 7 13:54:46 2018 -0800 enable multivalues insert (pandas-dev#19664) commit f33e84c Author: Ksenia <bobrovaksenia@gmail.com> Date: Wed Mar 7 22:09:42 2018 +0100 Moving tests in series/indexing to fixtures (pandas-dev#20014.1) (pandas-dev#20034) commit 2532a49 Author: Liam3851 <david.krych@gmail.com> Date: Wed Mar 7 13:04:22 2018 -0500 BUG: Fixes to msgpack support. (pandas-dev#19975) commit fd010de Author: Guilherme Beltramini <guilherme.beltramini@nubank.com.br> Date: Wed Mar 7 11:33:09 2018 -0300 to_sql also accepts Series (pandas-dev#20004) commit 8d462ed Author: Paul Reidy <paul_reidy@outlook.com> Date: Wed Mar 7 14:32:12 2018 +0000 EHN: Implement method argument for DataFrame.replace (pandas-dev#19894) commit d14fae8 Author: jbrockmendel <jbrockmendel@gmail.com> Date: Wed Mar 7 06:19:21 2018 -0800 cleanup ops (pandas-dev#19972) commit 776f2be Author: William Ayd <william.ayd@icloud.com> Date: Wed Mar 7 05:59:39 2018 -0800 Added .pytest_cache to gitignore (pandas-dev#20021) commit 460941f Author: jschendel <jschendel@users.noreply.github.com> Date: Wed Mar 7 06:57:51 2018 -0700 Fix typos in test_interval_new (pandas-dev#20026) commit 5782ab8 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Wed Mar 7 14:57:17 2018 +0100 DOC: enable matplotlib plot_directive to include figures in docstrings (pandas-dev#20015) commit dd2b224 Author: DataOmbudsman <DataOmbudsman@users.noreply.github.com> Date: Wed Mar 7 14:56:49 2018 +0100 DOC: updating docstring of Index.shift (pandas-dev#19996) commit 09c416c Author: William Ayd <william.ayd@icloud.com> Date: Wed Mar 7 05:56:16 2018 -0800 DOC: Updated kurt docstring (for pandas sprint) (pandas-dev#19999) commit ad15f80 Author: Kate Surta <kate.surta@gmail.com> Date: Wed Mar 7 16:55:48 2018 +0300 TST: Fix wrong argument in TestDataFrameAlterAxes.test_set_index_dst (pandas-dev#20019) commit f6ee9ac Author: Jeff Reback <jeff@reback.net> Date: Wed Mar 7 08:55:33 2018 -0500 TST: xfail clip tests under numpy-dev (pandas-dev#20035) xref pandas-dev#19976 commit 397e296 Author: Jeff Reback <jeff@reback.net> Date: Wed Mar 7 08:15:49 2018 -0500 TST: xfail some tests for mpl 2.2 compat (pandas-dev#20033) xref pandas-dev#20031 commit 56939b4 Author: luzpaz <luzpaz@users.noreply.github.com> Date: Wed Mar 7 06:10:39 2018 -0500 DOC: misc typos (pandas-dev#20029) commit 01b91c2 Author: alinde1 <32714875+alinde1@users.noreply.github.com> Date: Tue Mar 6 22:47:45 2018 +0100 DOC: is confusing for ddof parameter of sem, var and std functions (pandas-dev#19986) commit db82165 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Tue Mar 6 22:42:41 2018 +0100 CLN/DOC: cache_readonly: remove allow_setting + preserve docstring (pandas-dev#19991) commit e02f737 Author: Tom Augspurger <TomAugspurger@users.noreply.github.com> Date: Tue Mar 6 09:38:32 2018 -0600 DOC: add doc on ExtensionArray and extending pandas (pandas-dev#19936) commit 0ca77b3 Author: jbrockmendel <jbrockmendel@gmail.com> Date: Tue Mar 6 04:27:21 2018 -0800 Datetimelike add/sub catch cases more explicitly, tests (pandas-dev#19912) commit 0038bad Author: Matthew Roeschke <emailformattr@gmail.com> Date: Tue Mar 6 04:25:55 2018 -0800 month_name/day_name warnings followup (pandas-dev#20010) commit fd63c90 Author: Ksenia <bobrovaksenia@gmail.com> Date: Tue Mar 6 13:25:37 2018 +0100 TST: split series/test_indexing.py (pandas-dev#18614) (pandas-dev#20006) commit 6366bf0 Author: Jeff Reback <jeff@reback.net> Date: Tue Mar 6 07:25:17 2018 -0500 TST: clean deprecation warnings for xref pandas-dev#19980 (pandas-dev#20013) xfail some mpl > 2.1.2 tests commit fe61299 Author: William Ayd <william.ayd@icloud.com> Date: Tue Mar 6 00:30:13 2018 -0800 DOC: fixed dynamic import mechanics of make.py (pandas-dev#20005) commit 8a084eb Author: Grant Smith <grantsmith@gmail.com> Date: Tue Mar 6 03:29:26 2018 -0500 CLN: deprecate the pandas.tseries.plotting.tsplot function (GH18627) (pandas-dev#19980) commit aedbd94 Author: Jeff Reback <jeff@reback.net> Date: Mon Mar 5 06:36:41 2018 -0500 TST: text correction, xref pandas-dev#19987 commit cbffd19 Author: Bhavesh Poddar <bhavesh13103507@gmail.com> Date: Mon Mar 5 06:34:59 2018 -0500 fixed pytest deprecation warning (pandas-dev#19987) commit 058a16c Author: Matthew Roeschke <emailformattr@gmail.com> Date: Mon Mar 5 03:23:49 2018 -0800 CLN: Use generators in builtin functions (pandas-dev#19989) commit 607910b Author: Matthew Roeschke <emailformattr@gmail.com> Date: Sun Mar 4 12:15:37 2018 -0800 Add month names (pandas-dev#18164) commit 2fad756 Author: jbrockmendel <jbrockmendel@gmail.com> Date: Sun Mar 4 12:00:39 2018 -0800 transition period_helper to use pandas_datetimestruct (pandas-dev#19918) commit 53606ff Author: Liam3851 <david.krych@gmail.com> Date: Sun Mar 4 14:58:22 2018 -0500 BUG: Compat for pre-0.20 TimedeltaIndex and Float64Index pickles pandas-dev#19939 (pandas-dev#19943) commit 0bfb61b Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Fri Mar 2 22:35:45 2018 +0100 DOC: small updates to make.py script (pandas-dev#19951) * enable passing verbosity flag to sphinx * alias api for api.rst commit d1f3689 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Fri Mar 2 22:33:48 2018 +0100 DOC: fix some sphinx syntax warnings (pandas-dev#19962) commit 49f09cc Author: Tom Augspurger <TomAugspurger@users.noreply.github.com> Date: Fri Mar 2 15:20:28 2018 -0600 API: Added ExtensionArray constructor from scalars (pandas-dev#19913) commit d30d165 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Fri Mar 2 22:18:10 2018 +0100 DOC: update docstring validation script (pandas-dev#19960) commit a7a7f8c Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Fri Mar 2 13:49:59 2018 +0100 DOC: clarify version of ActivePython that includes pandas (pandas-dev#19964) commit b167483 Author: Gina <Dr-G@users.noreply.github.com> Date: Fri Mar 2 05:33:49 2018 -0600 DOC: update install.rst to include ActivePython distribution (pandas-dev#19908) commit e6c7dea Author: topper-123 <terji78@gmail.com> Date: Fri Mar 2 11:19:07 2018 +0000 ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part III) (pandas-dev#19884) commit d615f86 Author: Marc Garcia <garcia.marc@gmail.com> Date: Fri Mar 2 09:39:45 2018 +0000 DOC: Adding script to validate docstrings, and generate list of all functions/methods with state (pandas-dev#19898) commit 5f271eb Author: Yian <yian.shang@gmail.com> Date: Fri Mar 2 00:13:58 2018 +0100 BUG: Adding skipna as an option to groupby cumsum and cumprod (pandas-dev#19914) commit 072545d Author: David C Hall <davidchall@users.noreply.github.com> Date: Thu Mar 1 15:06:20 2018 -0800 ENH: Add option to disable MathJax (pandas-dev#19824). (pandas-dev#19856) commit d44a6ec Author: Yian <yian.shang@gmail.com> Date: Fri Mar 2 00:02:31 2018 +0100 Making to_datetime('today') and Timestamp('today') consistent (pandas-dev#19937) commit 87fefe2 Author: jbrockmendel <jbrockmendel@gmail.com> Date: Thu Mar 1 14:54:42 2018 -0800 dispatch Series[datetime64] comparison ops to DatetimeIndex (pandas-dev#19800) commit 9242248 Author: Matthew Roeschke <emailformattr@gmail.com> Date: Thu Mar 1 14:50:35 2018 -0800 BUG: DataFrame.diff(axis=0) with DatetimeTZ data (pandas-dev#19773) commit c5a1ef1 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Thu Mar 1 22:48:39 2018 +0100 DOC: remove empty attribute/method lists from class docstrings html page (pandas-dev#19949) commit 9958ce6 Author: jschendel <jschendel@users.noreply.github.com> Date: Thu Mar 1 04:14:19 2018 -0700 BUG: Preserve column metadata with DataFrame.astype (pandas-dev#19948) commit 3b4eb8d Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Thu Mar 1 12:12:35 2018 +0100 CLN: remove redundant clean_fill_method calls (pandas-dev#19947) commit c8859b5 Author: Joris Van den Bossche <jorisvandenbossche@gmail.com> Date: Thu Mar 1 10:35:05 2018 +0100 DOC: script to build single docstring page (pandas-dev#19840) commit 52559f5 Author: Matthew Roeschke <emailformattr@gmail.com> Date: Wed Feb 28 17:32:24 2018 -0800 ENH: Allow Timestamp to accept Nanosecond argument (pandas-dev#19889) commit 4a27697 Author: William Ayd <william.ayd@icloud.com> Date: Wed Feb 28 17:30:18 2018 -0800 Cythonized GroupBy any (pandas-dev#19722) commit 96b8bb1 Author: jschendel <jschendel@users.noreply.github.com> Date: Wed Feb 28 18:07:15 2018 -0700 ENH: Implement DataFrame.astype('category') (pandas-dev#18099) commit 6ef4be3 Author: Liam3851 <david.krych@gmail.com> Date: Wed Feb 28 06:14:11 2018 -0500 ENH: Allow literal (non-regex) replacement using .str.replace pandas-dev#16808 (pandas-dev#19584) commit 318a287 Author: README Bot <35302948+codetriage-readme-bot@users.noreply.github.com> Date: Wed Feb 28 05:07:28 2018 -0600 Add CodeTriage badge to pandas-dev/pandas (pandas-dev#19928) Adds a badge showing the number of people helping this repo on CodeTriage. commit 14a38a6 Author: Chris Catalfo <ccatalfo@users.noreply.github.com> Date: Wed Feb 28 03:14:23 2018 -0500 DOC: fixes pipe example in basics.rst due to statsmodel changes (pandas-dev#19923) commit dfe9d4a Author: Phil Ngo <ngo.phil@gmail.com> Date: Wed Feb 28 00:05:56 2018 -0800 DOC: fix Series.reset_index example (pandas-dev#19930) commit 9bdc5c8 Author: William Ayd <william.ayd@icloud.com> Date: Tue Feb 27 16:16:48 2018 -0800 Consistent Timedelta Writing for all Excel Engines (pandas-dev#19921) commit 61211a8 Author: jbrockmendel <jbrockmendel@gmail.com> Date: Tue Feb 27 16:11:47 2018 -0800 Assorted _libs cleanups (pandas-dev#19887)

topper-123 changed the title ~~initialization from dicts for py>=3.6 maintains insertion order~~ ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part III) Feb 24, 2018

jreback requested changes Feb 24, 2018

View reviewed changes

jreback added Reshaping Concat, Merge/Join, Stack/Unstack, Explode Compat pandas objects compatability with Numpy or Python functions labels Feb 24, 2018

topper-123 force-pushed the py36_dict_III branch from 0eeef4d to 4e57253 Compare February 24, 2018 19:05

jschendel reviewed Feb 24, 2018

View reviewed changes

topper-123 force-pushed the py36_dict_III branch from 4e57253 to 06d3c33 Compare February 25, 2018 09:24

topper-123 force-pushed the py36_dict_III branch 5 times, most recently from fa63f8d to 222e16c Compare February 25, 2018 12:22

jreback requested changes Feb 25, 2018

View reviewed changes

topper-123 force-pushed the py36_dict_III branch 2 times, most recently from 964fb29 to ffd2e18 Compare February 28, 2018 17:45

TomAugspurger reviewed Feb 28, 2018

View reviewed changes

topper-123 force-pushed the py36_dict_III branch from ffd2e18 to 410597a Compare March 1, 2018 06:00

topper-123 force-pushed the py36_dict_III branch from 410597a to 0fb9f66 Compare March 1, 2018 07:44

topper-123 force-pushed the py36_dict_III branch from 0fb9f66 to 820b0d3 Compare March 1, 2018 12:47

topper-123 closed this Mar 1, 2018

topper-123 reopened this Mar 1, 2018

TomAugspurger approved these changes Mar 1, 2018

View reviewed changes

jorisvandenbossche reviewed Mar 1, 2018

View reviewed changes

initialization from dicts for py>=3.6 maintains insertion order

778f08d

topper-123 force-pushed the py36_dict_III branch from 820b0d3 to 778f08d Compare March 2, 2018 08:28

jorisvandenbossche approved these changes Mar 2, 2018

View reviewed changes

jreback added this to the 0.23.0 milestone Mar 2, 2018

jreback added 2 commits March 2, 2018 06:12

Merge branch 'master' into PR_TOOL_MERGE_PR_19884

8ae2e8c

doc

473f564

jreback approved these changes Mar 2, 2018

View reviewed changes

jreback merged commit e6c7dea into pandas-dev:master Mar 2, 2018

topper-123 deleted the py36_dict_III branch March 2, 2018 12:14

wuhaochen mentioned this pull request May 16, 2018

change in pandas breaks doctest in python 3.6 networkx/networkx#2977

Closed

WillAyd mentioned this pull request Jun 17, 2018

Maintain Dict Ordering with Concat #21512

Closed

4 tasks

chris-b1 mentioned this pull request Aug 7, 2018

BUG: Regression creating DataFrame from nested dict #22227

Closed

		@@ -241,6 +241,17 @@ def test_constructor_dict(self):
		# Corner cases
		assert len(DataFrame({})) == 0

		@@ -775,6 +775,17 @@ def test_constructor_dict(self):
		expected = Series([1, 2, nan, 0], index=['b', 'c', 'd', 'a'])
		assert_series_equal(result, expected)


		Note the series above is ordered alphabetically by the index values.

		New behaviour (for Python >= 3.6):

ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part III) #19884

ENH: Let initialisation from dicts use insertion order for python >= 3.6 (part III) #19884

Conversation

topper-123 commented Feb 24, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

topper-123 commented Feb 24, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jschendel Feb 24, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

topper-123 commented Feb 25, 2018

codecov bot commented Feb 25, 2018 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

topper-123 Feb 28, 2018 • edited Loading

Choose a reason for hiding this comment

pep8speaks commented Feb 28, 2018 • edited Loading

Comment last updated on March 02, 2018 at 11:18 Hours UTC

Choose a reason for hiding this comment

Choose a reason for hiding this comment

topper-123 commented Mar 1, 2018 • edited Loading

topper-123 commented Mar 1, 2018

jorisvandenbossche left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

topper-123 commented Mar 2, 2018

jorisvandenbossche commented Mar 2, 2018

jorisvandenbossche left a comment

Choose a reason for hiding this comment

jreback commented Mar 2, 2018

topper-123 commented Feb 24, 2018 •

edited

Loading

jschendel Feb 24, 2018 •

edited

Loading

codecov bot commented Feb 25, 2018 •

edited

Loading

topper-123 Feb 28, 2018 •

edited

Loading

pep8speaks commented Feb 28, 2018 •

edited

Loading

topper-123 commented Mar 1, 2018 •

edited

Loading