Merge remote-tracking branch 'upstream/master' into boolean-array-kleene

Author: TomAugspurger

README.md

@@ -164,12 +164,11 @@ pip install pandas
 ```
 
 ## Dependencies
-- [NumPy](https://www.numpy.org): 1.13.3 or higher
-- [python-dateutil](https://labix.org/python-dateutil): 2.5.0 or higher
-- [pytz](https://pythonhosted.org/pytz): 2015.4 or higher
+- [NumPy](https://www.numpy.org)
+- [python-dateutil](https://labix.org/python-dateutil)
+- [pytz](https://pythonhosted.org/pytz)
 
-See the [full installation instructions](https://pandas.pydata.org/pandas-docs/stable/install.html#dependencies)
-for recommended and optional dependencies.
+See the [full installation instructions](https://pandas.pydata.org/pandas-docs/stable/install.html#dependencies) for minimum supported versions of required, recommended and optional dependencies.
 
 ## Installation from sources
 To install pandas from source you need Cython in addition to the normal

azure-pipelines.yml

@@ -16,95 +16,6 @@ jobs:
     name: Windows
     vmImage: vs2017-win2016
 
-- job: 'Checks'
-  pool:
-    vmImage: ubuntu-16.04
-  timeoutInMinutes: 90
-  steps:
-  - script: |
-      echo '##vso[task.prependpath]$(HOME)/miniconda3/bin'
-      echo '##vso[task.setvariable variable=ENV_FILE]environment.yml'
-      echo '##vso[task.setvariable variable=AZURE]true'
-    displayName: 'Setting environment variables'
-
-  # Do not require a conda environment
-  - script: ci/code_checks.sh patterns
-    displayName: 'Looking for unwanted patterns'
-    condition: true
-
-  - script: |
-      sudo apt-get update
-      sudo apt-get install -y libc6-dev-i386
-      ci/setup_env.sh
-    displayName: 'Setup environment and build pandas'
-    condition: true
-
-  # Do not require pandas
-  - script: |
-      source activate pandas-dev
-      ci/code_checks.sh lint
-    displayName: 'Linting'
-    condition: true
-
-  - script: |
-      source activate pandas-dev
-      ci/code_checks.sh dependencies
-    displayName: 'Dependencies consistency'
-    condition: true
-
-  # Require pandas
-  - script: |
-      source activate pandas-dev
-      ci/code_checks.sh code
-    displayName: 'Checks on imported code'
-    condition: true
-
-  - script: |
-      source activate pandas-dev
-      ci/code_checks.sh doctests
-    displayName: 'Running doctests'
-    condition: true
-
-  - script: |
-      source activate pandas-dev
-      ci/code_checks.sh docstrings
-    displayName: 'Docstring validation'
-    condition: true
-
-  - script: |
-      source activate pandas-dev
-      ci/code_checks.sh typing
-    displayName: 'Typing validation'
-    condition: true
-
-  - script: |
-      source activate pandas-dev
-      pytest --capture=no --strict scripts
-    displayName: 'Testing docstring validation script'
-    condition: true
-
-  - script: |
-      source activate pandas-dev
-      cd asv_bench
-      asv check -E existing
-      git remote add upstream https://github.com/pandas-dev/pandas.git
-      git fetch upstream
-      if git diff upstream/master --name-only | grep -q "^asv_bench/"; then
-          asv machine --yes
-          ASV_OUTPUT="$(asv dev)"
-          if [[ $(echo "$ASV_OUTPUT" | grep "failed") ]]; then
-              echo "##vso[task.logissue type=error]Benchmarks run with errors"
-              echo "$ASV_OUTPUT"
-              exit 1
-          else
-              echo "Benchmarks run without errors"
-          fi
-      else
-          echo "Benchmarks did not run, no changes detected"
-      fi
-    displayName: 'Running benchmarks'
-    condition: true
-
 - job: 'Web_and_Docs'
   pool:
     vmImage: ubuntu-16.04

doc/source/user_guide/scale.rst

@@ -94,7 +94,7 @@ Use efficient datatypes
 
 The default pandas data types are not the most memory efficient. This is
 especially true for text data columns with relatively few unique values (commonly
-referred to as "low-cardinality" data). By using more efficient data types you
+referred to as "low-cardinality" data). By using more efficient data types, you
 can store larger datasets in memory.
 
 .. ipython:: python

doc/source/whatsnew/v1.0.0.rst

@@ -403,6 +403,7 @@ or ``matplotlib.Axes.plot``. See :ref:`plotting.formatters` for more.
 
 - Floordiv of integer-dtyped array by :class:`Timedelta` now raises ``TypeError`` (:issue:`21036`)
 - Removed the previously deprecated :meth:`Index.summary` (:issue:`18217`)
+- Removed the previously deprecated "fastpath" keyword from the :class:`Index` constructor (:issue:`23110`)
 - Removed the previously deprecated :meth:`Series.get_value`, :meth:`Series.set_value`, :meth:`DataFrame.get_value`, :meth:`DataFrame.set_value` (:issue:`17739`)
 - Changed the the default value of `inplace` in :meth:`DataFrame.set_index` and :meth:`Series.set_axis`. It now defaults to False (:issue:`27600`)
 - Removed support for nested renaming in :meth:`DataFrame.aggregate`, :meth:`Series.aggregate`, :meth:`DataFrameGroupBy.aggregate`, :meth:`SeriesGroupBy.aggregate`, :meth:`Rolling.aggregate` (:issue:`18529`)

environment.yml

@@ -78,7 +78,7 @@ dependencies:
   - fastparquet>=0.3.2  # pandas.read_parquet, DataFrame.to_parquet
   - html5lib  # pandas.read_html
   - lxml  # pandas.read_html
-  - openpyxl  # pandas.read_excel, DataFrame.to_excel, pandas.ExcelWriter, pandas.ExcelFile
+  - openpyxl<=3.0.1  # pandas.read_excel, DataFrame.to_excel, pandas.ExcelWriter, pandas.ExcelFile
   - pyarrow>=0.13.1  # pandas.read_parquet, DataFrame.to_parquet, pandas.read_feather, DataFrame.to_feather
   - pyqt>=5.9.2  # pandas.read_clipboard
   - pytables>=3.4.2  # pandas.read_hdf, DataFrame.to_hdf

pandas/__init__.py

@@ -24,6 +24,7 @@
     _np_version_under1p15,
     _np_version_under1p16,
     _np_version_under1p17,
+    _np_version_under1p18,
 )
 
 try:

pandas/conftest.py

@@ -868,3 +868,16 @@ def float_frame():
     [30 rows x 4 columns]
     """
     return DataFrame(tm.getSeriesData())
+
+
+@pytest.fixture(params=[pd.Index, pd.Series], ids=["index", "series"])
+def index_or_series(request):
+    """
+    Fixture to parametrize over Index and Series, made necessary by a mypy
+    bug, giving an error:
+
+    List item 0 has incompatible type "Type[Series]"; expected "Type[PandasObject]"
+
+    See GH#?????
+    """
+    return request.param

pandas/core/indexes/base.py

@@ -265,14 +265,7 @@ def _outer_indexer(self, left, right):
     # Constructors
 
     def __new__(
-        cls,
-        data=None,
-        dtype=None,
-        copy=False,
-        name=None,
-        fastpath=None,
-        tupleize_cols=True,
-        **kwargs,
+        cls, data=None, dtype=None, copy=False, name=None, tupleize_cols=True, **kwargs,
     ) -> "Index":
 
         from .range import RangeIndex
@@ -284,16 +277,6 @@ def __new__(
         if name is None and hasattr(data, "name"):
             name = data.name
 
-        if fastpath is not None:
-            warnings.warn(
-                "The 'fastpath' keyword is deprecated, and will be "
-                "removed in a future version.",
-                FutureWarning,
-                stacklevel=2,
-            )
-            if fastpath:
-                return cls._simple_new(data, name)
-
         if isinstance(data, ABCPandasArray):
             # ensure users don't accidentally put a PandasArray in an index.
             data = data.to_numpy()

pandas/core/indexes/category.py

@@ -1,6 +1,5 @@
 import operator
 from typing import Any
-import warnings
 
 import numpy as np
 
@@ -172,19 +171,8 @@ def __new__(
         dtype=None,
         copy=False,
         name=None,
-        fastpath=None,
     ):
 
-        if fastpath is not None:
-            warnings.warn(
-                "The 'fastpath' keyword is deprecated, and will be "
-                "removed in a future version.",
-                FutureWarning,
-                stacklevel=2,
-            )
-            if fastpath:
-                return cls._simple_new(data, name=name, dtype=dtype)
-
         dtype = CategoricalDtype._from_values_or_dtype(data, categories, ordered, dtype)
 
         if name is None and hasattr(data, "name"):

pandas/core/indexes/datetimelike.py

@@ -284,7 +284,10 @@ def sort_values(self, return_indexer=False, ascending=True):
             sorted_index = self.take(_as)
             return sorted_index, _as
         else:
-            sorted_values = np.sort(self._ndarray_values)
+            # NB: using asi8 instead of _ndarray_values matters in numpy 1.18
+            #  because the treatment of NaT has been changed to put NaT last
+            #  instead of first.
+            sorted_values = np.sort(self.asi8)
             attribs = self._get_attributes_dict()
             freq = attribs["freq"]