Bug: adds support for unary plus #19297

deniederhut · 2018-01-18T13:02:38Z

closes Unary plus missing for Series and DataFrame #16073
tests added / passed
passes git diff upstream/master -u -- "*.py" | flake8 --diff
whatsnew entry

Adds missing unary plus operator from #16073. Adds typechecking behavior to unary neg, along with tests for both.

jreback

can you add a similar set of tests for Series?

jreback · 2018-01-18T23:41:21Z

doc/source/whatsnew/v0.23.0.txt

@@ -273,6 +273,7 @@ Build Changes
 Other API Changes
 ^^^^^^^^^^^^^^^^^

+- Unary ``+`` operator now permitted for ``Series`` and ``DataFrame`` (:issue:`16073`)


move to Other Enhancements, add the word numeric operations at the end.

jreback · 2018-01-18T23:42:11Z

pandas/core/generic.py

+
+    def __pos__(self):
+        values = _values_from_object(self)
+        if values.dtype == np.bool_:


is_bool_dtype (and can you change in __neg__ as well

jreback · 2018-01-18T23:42:33Z

pandas/tests/frame/test_operators.py

@@ -389,12 +389,42 @@ def test_logical_with_nas(self):
        assert_series_equal(result, expected)

    def test_neg(self):
-        # what to do?
-        assert_frame_equal(-self.frame, -1 * self.frame)
+        numeric = pd.DataFrame({


can you parameterize this

jreback · 2018-01-18T23:42:38Z

pandas/tests/frame/test_operators.py


    def test_invert(self):
        assert_frame_equal(-(self.frame < 0), ~(self.frame < 0))

+    def test_pos(self):


jreback · 2018-01-18T23:43:13Z

pandas/tests/frame/test_operators.py

+        assert_frame_equal(+numeric, +1 * numeric)
+        assert_frame_equal(+boolean, (+1 * boolean).astype(bool))
+        assert_series_equal(+timedelta, pd.to_timedelta(+1 * timedelta))
+        with pytest.raises(TypeError):


move the TypeError check (e.g. passing in invalid dtypes) to a separate test, and parameterize with all the invalid dtypes.

deniederhut · 2018-01-22T11:37:11Z

From the CI logs, it looks like the new operator trips some of the computational tests for other dtypes. I can clean these up, but would like to ask for some guidance on how we would like to handle operations on the timeperiod dtypes.

pep8speaks · 2018-01-22T11:41:40Z

Hello @deniederhut! Thanks for updating the PR.

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

Comment last updated on February 08, 2018 at 01:18 Hours UTC

jreback · 2018-01-22T11:42:03Z

@deniederhut

@jbrockmendel has been working a lot on the ops recently. what cases are in question? can you show (alternativley, you can xfail them and then point them out in your diff).

deniederhut · 2018-01-22T18:01:42Z

TestSeriesPeriod.test_ops_series_period and TestFramePeriod.test_ops_frame_period both had unary negatives which are currently not allowed for TimePeriod objects, such as:

tm.assert_series_equal(s - p, -exp)

Do we want to allow these?

TestEvalPythonPython.test_frame_pos and TestEvalPythonPython.test_series_pos are both testing to ensure that a unary plus always raises, even with numeric data. My assumption is that these would be okay to remove, since they are testing for the nonexistence of the feature this PR implements.

deniederhut · 2018-01-25T13:21:13Z

@jreback would it be better to defer the decision about TimePeriods, and xfail the problematic tests for now?

jreback · 2018-01-27T01:16:46Z

yes, ok with fixing Timedelta / periods after. pls put in the appopriate xfails. and rebase.

deniederhut · 2018-02-05T02:10:57Z

@jreback It's not clear to me why the Travis build is failing -- there are some warnings about IPython versions and SparseBlocks, but no failed tests. Is there something obvious that I am missing?

jreback · 2018-02-05T02:19:46Z

you are failing linting

see the very end

deniederhut · 2018-02-05T12:56:05Z

🤦‍♀️ yup there it is. Thanks!

codecov · 2018-02-05T12:56:37Z

Codecov Report

Merging #19297 into master will increase coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master   #19297      +/-   ##
==========================================
+ Coverage   91.62%   91.62%   +<.01%     
==========================================
  Files         150      150              
  Lines       48773    48790      +17     
==========================================
+ Hits        44687    44704      +17     
  Misses       4086     4086

Flag	Coverage Δ
#multiple	`89.99% <100%> (ø)`	⬆️
#single	`41.73% <9.09%> (-0.02%)`	⬇️

Impacted Files	Coverage Δ
pandas/core/generic.py	`95.93% <100%> (+0.01%)`	⬆️
pandas/core/indexes/category.py	`97.26% <0%> (-0.26%)`	⬇️
pandas/core/reshape/merge.py	`94.25% <0%> (+0.03%)`	⬆️
pandas/core/arrays/categorical.py	`94.87% <0%> (+0.13%)`	⬆️

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 affb5d9...2b8072e. Read the comment docs.

jreback · 2018-02-06T01:18:25Z

pandas/tests/frame/test_arithmetic.py

@@ -111,6 +111,8 @@ def test_df_radd_str(self):

 class TestPeriodFrameArithmetic(object):

+    @pytest.mark.xfail(reason="Unary negative is currently not permitted "


@deniederhut so period - period is now going to raise? this is #13077.

I am not actually sure what we should do here. I think we can re-enable these though, at least for now.

(you can use is_period_arraylike to detect this)

The test is marked xfail because of the -exp in this assertion tm.assert_frame_equal(df - p, -exp), not the initial subtraction. Happy to allow it and unmark the tests.

Should the expected return values in the tests be scalars? I think period - period currently returns an integer.

i c
the return value should be integers

Adds special method for unary plus to ndframe generics, and restricts the allowed dtypes for the unary negative. Closes pandas-dev#16073.

jreback · 2018-02-08T11:31:34Z

pandas/tests/frame/test_operators.py

+            pd.DataFrame({'a': pd.Series(pd.to_timedelta([1, -1]))}))
+    ])
+    def test_neg_numeric(self, df, expected):
+        assert_frame_equal(-df, expected)


@jbrockmendel I think on your list is to rationalize locations of arithmetic tests between test_arithmetic and test_operations

jreback · 2018-02-08T11:32:25Z

thanks @deniederhut nice patch! sometimes little issues can morph, nice job!

deniederhut · 2018-02-08T12:53:18Z

Thanks so much for your help @jreback !

jreback requested changes Jan 18, 2018

View reviewed changes

jreback added Enhancement Dtype Conversions Unexpected or buggy dtype conversions Numeric Operations Arithmetic, Comparison, and Logical operations labels Jan 18, 2018

deniederhut force-pushed the bug/unary-plus branch from c61a108 to 047f62e Compare January 23, 2018 01:28

deniederhut force-pushed the bug/unary-plus branch 2 times, most recently from 7145fe4 to 3d0844a Compare February 2, 2018 11:54

deniederhut force-pushed the bug/unary-plus branch from 3d0844a to 1bcd080 Compare February 5, 2018 12:56

jreback requested changes Feb 6, 2018

View reviewed changes

deniederhut added 10 commits February 7, 2018 09:10

BUG: correct behavior for unary plus and negative

b018715

Adds special method for unary plus to ndframe generics, and restricts the allowed dtypes for the unary negative. Closes pandas-dev#16073.

TST: adds tests for unary plus and unary neg

8429618

DOC: adds whatsnew entry for unary plus operator

bfe628a

CLN: fixes bracket indents

6c0d456

DOC: language changes to whatsnew

3e887c2

BUG: replaces typechecks with is_bool_dtype

f88d38d

TST: paramterizes pos and neg tests

c18a976

CLN: pep8 fixes

a696d64

BUG: fixes NameError introduced by rebase

122fcd6

TST: removes expected raises for unary pos from eval tests

ff8b250

deniederhut force-pushed the bug/unary-plus branch from 1bcd080 to 5203027 Compare February 7, 2018 15:10

deniederhut added 2 commits February 7, 2018 19:17

ENH: permits unary pos for period objects

11e150d

TST: fixes arithmetic tests for object output of period math

2b8072e

deniederhut force-pushed the bug/unary-plus branch from 5203027 to 2b8072e Compare February 8, 2018 01:18

jreback reviewed Feb 8, 2018

View reviewed changes

jreback added this to the 0.23.0 milestone Feb 8, 2018

jreback approved these changes Feb 8, 2018

View reviewed changes

jreback merged commit b835127 into pandas-dev:master Feb 8, 2018

harisbal pushed a commit to harisbal/pandas that referenced this pull request Feb 28, 2018

Bug: adds support for unary plus (pandas-dev#19297)

91c76cc

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bug: adds support for unary plus #19297

Bug: adds support for unary plus #19297

deniederhut commented Jan 18, 2018 •

edited

Loading

jreback left a comment

jreback Jan 18, 2018

jreback Jan 18, 2018

jreback Jan 18, 2018

jreback Jan 18, 2018

jreback Jan 18, 2018

deniederhut commented Jan 22, 2018

pep8speaks commented Jan 22, 2018 •

edited

Loading

jreback commented Jan 22, 2018

deniederhut commented Jan 22, 2018

deniederhut commented Jan 25, 2018

jreback commented Jan 27, 2018

deniederhut commented Feb 5, 2018

jreback commented Feb 5, 2018

deniederhut commented Feb 5, 2018

codecov bot commented Feb 5, 2018 •

edited

Loading

jreback Feb 6, 2018

deniederhut Feb 6, 2018

jreback Feb 6, 2018

jreback Feb 8, 2018

jreback commented Feb 8, 2018

deniederhut commented Feb 8, 2018

		@@ -111,6 +111,8 @@ def test_df_radd_str(self):

		class TestPeriodFrameArithmetic(object):

		@pytest.mark.xfail(reason="Unary negative is currently not permitted "

Bug: adds support for unary plus #19297

Bug: adds support for unary plus #19297

Conversation

deniederhut commented Jan 18, 2018 • edited Loading

jreback left a 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

deniederhut commented Jan 22, 2018

pep8speaks commented Jan 22, 2018 • edited Loading

Comment last updated on February 08, 2018 at 01:18 Hours UTC

jreback commented Jan 22, 2018

deniederhut commented Jan 22, 2018

deniederhut commented Jan 25, 2018

jreback commented Jan 27, 2018

deniederhut commented Feb 5, 2018

jreback commented Feb 5, 2018

deniederhut commented Feb 5, 2018

codecov bot commented Feb 5, 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

jreback commented Feb 8, 2018

deniederhut commented Feb 8, 2018

deniederhut commented Jan 18, 2018 •

edited

Loading

pep8speaks commented Jan 22, 2018 •

edited

Loading

codecov bot commented Feb 5, 2018 •

edited

Loading