Add files using upload-large-folder tool

lib/python3.12/site-packages/pandas/pyproject.toml

@@ -0,0 +1,811 @@
+[build-system]
+# Minimum requirements for the build system to execute.
+# See https://github.com/scipy/scipy/pull/12940 for the AIX issue.
+requires = [
+    "meson-python==0.13.1",
+    "meson==1.2.1",
+    "wheel",
+    "Cython~=3.0.5",  # Note: sync with setup.py, environment.yml and asv.conf.json
+    # Force numpy higher than 2.0, so that built wheels are compatible
+    # with both numpy 1 and 2
+    "numpy>=2.0",
+    "versioneer[toml]"
+]
+
+build-backend = "mesonpy"
+
+[project]
+name = 'pandas'
+dynamic = [
+  'version'
+]
+description = 'Powerful data structures for data analysis, time series, and statistics'
+readme = 'README.md'
+authors = [
+  { name = 'The Pandas Development Team', email='pandas-dev@python.org' },
+]
+license = {file = 'LICENSE'}
+requires-python = '>=3.9'
+dependencies = [
+  "numpy>=1.22.4; python_version<'3.11'",
+  "numpy>=1.23.2; python_version=='3.11'",
+  "numpy>=1.26.0; python_version>='3.12'",
+  "python-dateutil>=2.8.2",
+  "pytz>=2020.1",
+  "tzdata>=2022.7"
+]
+classifiers = [
+    'Development Status :: 5 - Production/Stable',
+    'Environment :: Console',
+    'Intended Audience :: Science/Research',
+    'License :: OSI Approved :: BSD License',
+    'Operating System :: OS Independent',
+    'Programming Language :: Cython',
+    'Programming Language :: Python',
+    'Programming Language :: Python :: 3',
+    'Programming Language :: Python :: 3 :: Only',
+    'Programming Language :: Python :: 3.9',
+    'Programming Language :: Python :: 3.10',
+    'Programming Language :: Python :: 3.11',
+    'Programming Language :: Python :: 3.12',
+    'Topic :: Scientific/Engineering'
+]
+
+[project.urls]
+homepage = 'https://pandas.pydata.org'
+documentation = 'https://pandas.pydata.org/docs/'
+repository = 'https://github.com/pandas-dev/pandas'
+
+[project.entry-points."pandas_plotting_backends"]
+matplotlib = "pandas:plotting._matplotlib"
+
+[project.optional-dependencies]
+test = ['hypothesis>=6.46.1', 'pytest>=7.3.2', 'pytest-xdist>=2.2.0']
+pyarrow = ['pyarrow>=10.0.1']
+performance = ['bottleneck>=1.3.6', 'numba>=0.56.4', 'numexpr>=2.8.4']
+computation = ['scipy>=1.10.0', 'xarray>=2022.12.0']
+fss = ['fsspec>=2022.11.0']
+aws = ['s3fs>=2022.11.0']
+gcp = ['gcsfs>=2022.11.0', 'pandas-gbq>=0.19.0']
+excel = ['odfpy>=1.4.1', 'openpyxl>=3.1.0', 'python-calamine>=0.1.7', 'pyxlsb>=1.0.10', 'xlrd>=2.0.1', 'xlsxwriter>=3.0.5']
+parquet = ['pyarrow>=10.0.1']
+feather = ['pyarrow>=10.0.1']
+hdf5 = [# blosc only available on conda (https://github.com/Blosc/python-blosc/issues/297)
+        #'blosc>=1.20.1',
+        'tables>=3.8.0']
+spss = ['pyreadstat>=1.2.0']
+postgresql = ['SQLAlchemy>=2.0.0', 'psycopg2>=2.9.6', 'adbc-driver-postgresql>=0.8.0']
+mysql = ['SQLAlchemy>=2.0.0', 'pymysql>=1.0.2']
+sql-other = ['SQLAlchemy>=2.0.0', 'adbc-driver-postgresql>=0.8.0', 'adbc-driver-sqlite>=0.8.0']
+html = ['beautifulsoup4>=4.11.2', 'html5lib>=1.1', 'lxml>=4.9.2']
+xml = ['lxml>=4.9.2']
+plot = ['matplotlib>=3.6.3']
+output-formatting = ['jinja2>=3.1.2', 'tabulate>=0.9.0']
+clipboard = ['PyQt5>=5.15.9', 'qtpy>=2.3.0']
+compression = ['zstandard>=0.19.0']
+consortium-standard = ['dataframe-api-compat>=0.1.7']
+all = ['adbc-driver-postgresql>=0.8.0',
+       'adbc-driver-sqlite>=0.8.0',
+       'beautifulsoup4>=4.11.2',
+       # blosc only available on conda (https://github.com/Blosc/python-blosc/issues/297)
+       #'blosc>=1.21.3',
+       'bottleneck>=1.3.6',
+       'dataframe-api-compat>=0.1.7',
+       'fastparquet>=2022.12.0',
+       'fsspec>=2022.11.0',
+       'gcsfs>=2022.11.0',
+       'html5lib>=1.1',
+       'hypothesis>=6.46.1',
+       'jinja2>=3.1.2',
+       'lxml>=4.9.2',
+       'matplotlib>=3.6.3',
+       'numba>=0.56.4',
+       'numexpr>=2.8.4',
+       'odfpy>=1.4.1',
+       'openpyxl>=3.1.0',
+       'pandas-gbq>=0.19.0',
+       'psycopg2>=2.9.6',
+       'pyarrow>=10.0.1',
+       'pymysql>=1.0.2',
+       'PyQt5>=5.15.9',
+       'pyreadstat>=1.2.0',
+       'pytest>=7.3.2',
+       'pytest-xdist>=2.2.0',
+       'python-calamine>=0.1.7',
+       'pyxlsb>=1.0.10',
+       'qtpy>=2.3.0',
+       'scipy>=1.10.0',
+       's3fs>=2022.11.0',
+       'SQLAlchemy>=2.0.0',
+       'tables>=3.8.0',
+       'tabulate>=0.9.0',
+       'xarray>=2022.12.0',
+       'xlrd>=2.0.1',
+       'xlsxwriter>=3.0.5',
+       'zstandard>=0.19.0']
+
+# TODO: Remove after setuptools support is dropped.
+[tool.setuptools]
+include-package-data = true
+
+[tool.setuptools.packages.find]
+include = ["pandas", "pandas.*"]
+namespaces = false
+
+[tool.setuptools.exclude-package-data]
+"*" = ["*.c", "*.h"]
+
+# See the docstring in versioneer.py for instructions. Note that you must
+# re-run 'versioneer.py setup' after changing this section, and commit the
+# resulting files.
+[tool.versioneer]
+VCS = "git"
+style = "pep440"
+versionfile_source = "pandas/_version.py"
+versionfile_build = "pandas/_version.py"
+tag_prefix = "v"
+parentdir_prefix = "pandas-"
+
+[tool.meson-python.args]
+setup = ['--vsenv'] # For Windows
+
+[tool.cibuildwheel]
+skip = "cp36-* cp37-* cp38-* pp* *_i686 *_ppc64le *_s390x"
+build-verbosity = "3"
+environment = {LDFLAGS="-Wl,--strip-all"}
+# pytz 2024.2 causing some failures
+test-requires = "hypothesis>=6.46.1 pytest>=7.3.2 pytest-xdist>=2.2.0 pytz<2024.2"
+test-command = """
+  PANDAS_CI='1' python -c 'import pandas as pd; \
+  pd.test(extra_args=["-m not clipboard and not single_cpu and not slow and not network and not db", "-n 2", "--no-strict-data-files"]); \
+  pd.test(extra_args=["-m not clipboard and single_cpu and not slow and not network and not db", "--no-strict-data-files"]);' \
+  """
+free-threaded-support = true
+before-build = "PACKAGE_DIR={package} bash {package}/scripts/cibw_before_build.sh"
+
+[tool.cibuildwheel.windows]
+before-build = "pip install delvewheel && bash {package}/scripts/cibw_before_build.sh"
+repair-wheel-command = "delvewheel repair -w {dest_dir} {wheel}"
+
+[[tool.cibuildwheel.overrides]]
+select = "*-manylinux_aarch64*"
+test-command = """
+  PANDAS_CI='1' python -c 'import pandas as pd; \
+  pd.test(extra_args=["-m not clipboard and not single_cpu and not slow and not network and not db and not fails_arm_wheels", "-n 2", "--no-strict-data-files"]); \
+  pd.test(extra_args=["-m not clipboard and single_cpu and not slow and not network and not db", "--no-strict-data-files"]);' \
+  """
+
+[[tool.cibuildwheel.overrides]]
+select = "*-musllinux*"
+before-test = "apk update && apk add musl-locales"
+
+[[tool.cibuildwheel.overrides]]
+select = "*-win*"
+# We test separately for Windows, since we use
+# the windowsservercore docker image to check if any dlls are
+# missing from the wheel
+test-command = ""
+
+[[tool.cibuildwheel.overrides]]
+# Don't strip wheels on macOS.
+# macOS doesn't support stripping wheels with linker
+# https://github.com/MacPython/numpy-wheels/pull/87#issuecomment-624878264
+select = "*-macosx*"
+environment = {CFLAGS="-g0"}
+
+[tool.black]
+target-version = ['py39', 'py310']
+required-version = '23.11.0'
+exclude = '''
+(
+    asv_bench/env
+  | \.egg
+  | \.git
+  | \.hg
+  | \.mypy_cache
+  | \.nox
+  | \.tox
+  | \.venv
+  | _build
+  | buck-out
+  | build
+  | dist
+  | setup.py
+)
+'''
+
+[tool.ruff]
+line-length = 88
+target-version = "py310"
+fix = true
+unfixable = []
+typing-modules = ["pandas._typing"]
+
+select = [
+  # pyflakes
+  "F",
+  # pycodestyle
+  "E", "W",
+  # flake8-2020
+  "YTT",
+  # flake8-bugbear
+  "B",
+  # flake8-quotes
+  "Q",
+  # flake8-debugger
+  "T10",
+  # flake8-gettext
+  "INT",
+  # pylint
+  "PL",
+  # misc lints
+  "PIE",
+  # flake8-pyi
+  "PYI",
+  # tidy imports
+  "TID",
+  # implicit string concatenation
+  "ISC",
+  # type-checking imports
+  "TCH",
+  # comprehensions
+  "C4",
+  # pygrep-hooks
+  "PGH",
+  # Ruff-specific rules
+  "RUF",
+  # flake8-bandit: exec-builtin
+  "S102",
+  # numpy-legacy-random
+  "NPY002",
+  # Perflint
+  "PERF",
+  # flynt
+  "FLY",
+  # flake8-logging-format
+  "G",
+  # flake8-future-annotations
+  "FA",
+]
+
+ignore = [
+  ### Intentionally disabled
+  # space before : (needed for how black formats slicing)
+  "E203",
+  # module level import not at top of file
+  "E402",
+  # do not assign a lambda expression, use a def
+  "E731",
+  # line break before binary operator
+  # "W503",  # not yet implemented
+  # line break after binary operator
+  # "W504",  # not yet implemented
+  # controversial
+  "B006",
+  # controversial
+  "B007",
+  # controversial
+  "B008",
+  # setattr is used to side-step mypy
+  "B009",
+  # getattr is used to side-step mypy
+  "B010",
+  # tests use assert False
+  "B011",
+  # tests use comparisons but not their returned value
+  "B015",
+  # false positives
+  "B019",
+  # Loop control variable overrides iterable it iterates
+  "B020",
+  # Function definition does not bind loop variable
+  "B023",
+  # Functions defined inside a loop must not use variables redefined in the loop
+  # "B301",  # not yet implemented
+  # Only works with python >=3.10
+  "B905",
+  # Too many arguments to function call
+  "PLR0913",
+  # Too many returns
+  "PLR0911",
+  # Too many branches
+  "PLR0912",
+  # Too many statements
+  "PLR0915",
+  # Redefined loop name
+  "PLW2901",
+  # Global statements are discouraged
+  "PLW0603",
+  # Docstrings should not be included in stubs
+  "PYI021",
+  # Use `typing.NamedTuple` instead of `collections.namedtuple`
+  "PYI024",
+  # No builtin `eval()` allowed
+  "PGH001",
+  # compare-to-empty-string
+  "PLC1901",
+  # while int | float can be shortened to float, the former is more explicit
+  "PYI041",
+  # incorrect-dict-iterator, flags valid Series.items usage
+  "PERF102",
+  # try-except-in-loop, becomes useless in Python 3.11
+  "PERF203",
+
+
+  ### TODO: Enable gradually
+  # Useless statement
+  "B018",
+  # Within an except clause, raise exceptions with ...
+  "B904",
+  # Magic number
+  "PLR2004",
+  # comparison-with-itself
+  "PLR0124",
+  # Consider `elif` instead of `else` then `if` to remove indentation level
+  "PLR5501",
+  # collection-literal-concatenation
+  "RUF005",
+  # pairwise-over-zipped (>=PY310 only)
+  "RUF007",
+  # explicit-f-string-type-conversion
+  "RUF010",
+  # mutable-class-default
+  "RUF012"
+]
+
+exclude = [
+  "doc/sphinxext/*.py",
+  "doc/build/*.py",
+  "doc/temp/*.py",
+  ".eggs/*.py",
+  # vendored files
+  "pandas/util/version/*",
+  "pandas/io/clipboard/__init__.py",
+  # exclude asv benchmark environments from linting
+  "env",
+]
+
+[tool.ruff.per-file-ignores]
+# relative imports allowed for asv_bench
+"asv_bench/*" = ["TID", "NPY002"]
+# to be enabled gradually
+"pandas/core/*" = ["PLR5501"]
+"pandas/tests/*" = ["B028", "FLY"]
+"scripts/*" = ["B028"]
+# Keep this one enabled
+"pandas/_typing.py" = ["TCH"]
+
+[tool.pylint.messages_control]
+max-line-length = 88
+disable = [
+ # intentionally turned off
+  "bad-mcs-classmethod-argument",
+  "broad-except",
+  "c-extension-no-member",
+  "comparison-with-itself",
+  "consider-using-enumerate",
+  "import-error",
+  "import-outside-toplevel",
+  "invalid-name",
+  "invalid-unary-operand-type",
+  "line-too-long",
+  "no-else-continue",
+  "no-else-raise",
+  "no-else-return",
+  "no-member",
+  "no-name-in-module",
+  "not-an-iterable",
+  "overridden-final-method",
+  "pointless-statement",
+  "redundant-keyword-arg",
+  "singleton-comparison",
+  "too-many-ancestors",
+  "too-many-arguments",
+  "too-many-boolean-expressions",
+  "too-many-branches",
+  "too-many-function-args",
+  "too-many-instance-attributes",
+  "too-many-locals",
+  "too-many-nested-blocks",
+  "too-many-public-methods",
+  "too-many-return-statements",
+  "too-many-statements",
+  "unexpected-keyword-arg",
+  "ungrouped-imports",
+  "unsubscriptable-object",
+  "unsupported-assignment-operation",
+  "unsupported-membership-test",
+  "unused-import",
+  "use-dict-literal",
+  "use-implicit-booleaness-not-comparison",
+  "use-implicit-booleaness-not-len",
+  "wrong-import-order",
+  "wrong-import-position",
+  "redefined-loop-name",
+
+ # misc
+  "abstract-class-instantiated",
+  "no-value-for-parameter",
+  "undefined-variable",
+  "unpacking-non-sequence",
+  "used-before-assignment",
+
+ # pylint type "C": convention, for programming standard violation
+  "missing-class-docstring",
+  "missing-function-docstring",
+  "missing-module-docstring",
+  "superfluous-parens",
+  "too-many-lines",
+  "unidiomatic-typecheck",
+  "unnecessary-dunder-call",
+  "unnecessary-lambda-assignment",
+
+  # pylint type "R": refactor, for bad code smell
+  "consider-using-with",
+  "cyclic-import",
+  "duplicate-code",
+  "inconsistent-return-statements",
+  "redefined-argument-from-local",
+  "too-few-public-methods",
+
+  # pylint type "W": warning, for python specific problems
+  "abstract-method",
+  "arguments-differ",
+  "arguments-out-of-order",
+  "arguments-renamed",
+  "attribute-defined-outside-init",
+  "broad-exception-raised",
+  "comparison-with-callable",
+  "dangerous-default-value",
+  "deprecated-module",
+  "eval-used",
+  "expression-not-assigned",
+  "fixme",
+  "global-statement",
+  "invalid-overridden-method",
+  "keyword-arg-before-vararg",
+  "possibly-unused-variable",
+  "protected-access",
+  "raise-missing-from",
+  "redefined-builtin",
+  "redefined-outer-name",
+  "self-cls-assignment",
+  "signature-differs",
+  "super-init-not-called",
+  "try-except-raise",
+  "unnecessary-lambda",
+  "unused-argument",
+  "unused-variable",
+  "using-constant-test"
+]
+
+[tool.pytest.ini_options]
+# sync minversion with pyproject.toml & install.rst
+minversion = "7.3.2"
+addopts = "--strict-markers --strict-config --capture=no --durations=30 --junitxml=test-data.xml"
+empty_parameter_set_mark = "fail_at_collect"
+xfail_strict = true
+testpaths = "pandas"
+doctest_optionflags = [
+  "NORMALIZE_WHITESPACE",
+  "IGNORE_EXCEPTION_DETAIL",
+  "ELLIPSIS",
+]
+filterwarnings = [
+  "error:::pandas",
+  "error::ResourceWarning",
+  "error::pytest.PytestUnraisableExceptionWarning",
+  # TODO(PY311-minimum): Specify EncodingWarning
+  # Ignore 3rd party EncodingWarning but raise on pandas'
+  "ignore:.*encoding.* argument not specified",
+  "error:.*encoding.* argument not specified::pandas",
+  "ignore:.*ssl.SSLSocket:pytest.PytestUnraisableExceptionWarning",
+  "ignore:.*ssl.SSLSocket:ResourceWarning",
+  # GH 44844: Can remove once minimum matplotlib version >= 3.7
+  "ignore:.*FileIO:pytest.PytestUnraisableExceptionWarning",
+  "ignore:.*BufferedRandom:ResourceWarning",
+  "ignore::ResourceWarning:asyncio",
+  # From plotting doctests
+  "ignore:More than 20 figures have been opened:RuntimeWarning",
+  # Will be fixed in numba 0.56: https://github.com/numba/numba/issues/7758
+  "ignore:`np.MachAr` is deprecated:DeprecationWarning:numba",
+  "ignore:.*urllib3:DeprecationWarning:botocore",
+  "ignore:Setuptools is replacing distutils.:UserWarning:_distutils_hack",
+  # https://github.com/PyTables/PyTables/issues/822
+  "ignore:a closed node found in the registry:UserWarning:tables",
+  "ignore:`np.object` is a deprecated:DeprecationWarning:tables",
+  "ignore:tostring:DeprecationWarning:tables",
+  "ignore:distutils Version classes are deprecated:DeprecationWarning:pandas_datareader",
+  "ignore:distutils Version classes are deprecated:DeprecationWarning:numexpr",
+  "ignore:distutils Version classes are deprecated:DeprecationWarning:fastparquet",
+  "ignore:distutils Version classes are deprecated:DeprecationWarning:fsspec",
+  # Can be removed once https://github.com/numpy/numpy/pull/24794 is merged
+  "ignore:.*In the future `np.long` will be defined as.*:FutureWarning",
+]
+junit_family = "xunit2"
+markers = [
+  "single_cpu: tests that should run on a single cpu only",
+  "slow: mark a test as slow",
+  "network: mark a test as network",
+  "db: tests requiring a database (mysql or postgres)",
+  "clipboard: mark a pd.read_clipboard test",
+  "arm_slow: mark a test as slow for arm64 architecture",
+  "skip_ubsan: Tests known to fail UBSAN check",
+  # TODO: someone should investigate this ...
+  # these tests only fail in the wheel builder and don't fail in regular
+  # ARM CI
+  "fails_arm_wheels: Tests that fail in the ARM wheel build only",
+]
+
+[tool.mypy]
+# Import discovery
+mypy_path = "typings"
+files = ["pandas", "typings"]
+namespace_packages = false
+explicit_package_bases = false
+ignore_missing_imports = true
+follow_imports = "normal"
+follow_imports_for_stubs = false
+no_site_packages = false
+no_silence_site_packages = false
+# Platform configuration
+python_version = "3.11"
+platform = "linux-64"
+# Disallow dynamic typing
+disallow_any_unimported = false # TODO
+disallow_any_expr = false # TODO
+disallow_any_decorated = false # TODO
+disallow_any_explicit = false # TODO
+disallow_any_generics = false # TODO
+disallow_subclassing_any = false # TODO
+# Untyped definitions and calls
+disallow_untyped_calls = true
+disallow_untyped_defs = true
+disallow_incomplete_defs = true
+check_untyped_defs = true
+disallow_untyped_decorators = true
+# None and Optional handling
+no_implicit_optional = true
+strict_optional = true
+# Configuring warnings
+warn_redundant_casts = true
+warn_unused_ignores = true
+warn_no_return = true
+warn_return_any = false # TODO
+warn_unreachable = false # GH#27396
+# Suppressing errors
+ignore_errors = false
+enable_error_code = "ignore-without-code"
+# Miscellaneous strictness flags
+allow_untyped_globals = false
+allow_redefinition = false
+local_partial_types = false
+implicit_reexport = true
+strict_equality = true
+# Configuring error messages
+show_error_context = false
+show_column_numbers = false
+show_error_codes = true
+
+[[tool.mypy.overrides]]
+module = [
+  "pandas._config.config", # TODO
+  "pandas._libs.*",
+  "pandas._testing.*", # TODO
+  "pandas.arrays", # TODO
+  "pandas.compat.numpy.function", # TODO
+  "pandas.compat._optional", # TODO
+  "pandas.compat.compressors", # TODO
+  "pandas.compat.pickle_compat", # TODO
+  "pandas.core._numba.executor", # TODO
+  "pandas.core.array_algos.datetimelike_accumulations", # TODO
+  "pandas.core.array_algos.masked_accumulations", # TODO
+  "pandas.core.array_algos.masked_reductions", # TODO
+  "pandas.core.array_algos.putmask", # TODO
+  "pandas.core.array_algos.quantile", # TODO
+  "pandas.core.array_algos.replace", # TODO
+  "pandas.core.array_algos.take", # TODO
+  "pandas.core.arrays.*", # TODO
+  "pandas.core.computation.*", # TODO
+  "pandas.core.dtypes.astype", # TODO
+  "pandas.core.dtypes.cast", # TODO
+  "pandas.core.dtypes.common", # TODO
+  "pandas.core.dtypes.concat", # TODO
+  "pandas.core.dtypes.dtypes", # TODO
+  "pandas.core.dtypes.generic", # TODO
+  "pandas.core.dtypes.inference", # TODO
+  "pandas.core.dtypes.missing", # TODO
+  "pandas.core.groupby.categorical", # TODO
+  "pandas.core.groupby.generic", # TODO
+  "pandas.core.groupby.grouper", # TODO
+  "pandas.core.groupby.groupby", # TODO
+  "pandas.core.groupby.ops", # TODO
+  "pandas.core.indexers.*", # TODO
+  "pandas.core.indexes.*", # TODO
+  "pandas.core.interchange.column", # TODO
+  "pandas.core.interchange.dataframe_protocol", # TODO
+  "pandas.core.interchange.from_dataframe", # TODO
+  "pandas.core.internals.*", # TODO
+  "pandas.core.methods.*", # TODO
+  "pandas.core.ops.array_ops", # TODO
+  "pandas.core.ops.common", # TODO
+  "pandas.core.ops.invalid", # TODO
+  "pandas.core.ops.mask_ops", # TODO
+  "pandas.core.ops.missing", # TODO
+  "pandas.core.reshape.*", # TODO
+  "pandas.core.strings.*", # TODO
+  "pandas.core.tools.*", # TODO
+  "pandas.core.window.common", # TODO
+  "pandas.core.window.ewm", # TODO
+  "pandas.core.window.expanding", # TODO
+  "pandas.core.window.numba_", # TODO
+  "pandas.core.window.online", # TODO
+  "pandas.core.window.rolling", # TODO
+  "pandas.core.accessor", # TODO
+  "pandas.core.algorithms", # TODO
+  "pandas.core.apply", # TODO
+  "pandas.core.arraylike", # TODO
+  "pandas.core.base", # TODO
+  "pandas.core.common", # TODO
+  "pandas.core.config_init", # TODO
+  "pandas.core.construction", # TODO
+  "pandas.core.flags", # TODO
+  "pandas.core.frame", # TODO
+  "pandas.core.generic", # TODO
+  "pandas.core.indexing", # TODO
+  "pandas.core.missing", # TODO
+  "pandas.core.nanops", # TODO
+  "pandas.core.resample", # TODO
+  "pandas.core.roperator", # TODO
+  "pandas.core.sample", # TODO
+  "pandas.core.series", # TODO
+  "pandas.core.sorting", # TODO
+  "pandas.errors", # TODO
+  "pandas.io.clipboard", # TODO
+  "pandas.io.excel._base", # TODO
+  "pandas.io.excel._odfreader", # TODO
+  "pandas.io.excel._odswriter", # TODO
+  "pandas.io.excel._openpyxl", # TODO
+  "pandas.io.excel._pyxlsb", # TODO
+  "pandas.io.excel._xlrd", # TODO
+  "pandas.io.excel._xlsxwriter", # TODO
+  "pandas.io.formats.console", # TODO
+  "pandas.io.formats.css", # TODO
+  "pandas.io.formats.excel", # TODO
+  "pandas.io.formats.format", # TODO
+  "pandas.io.formats.info", # TODO
+  "pandas.io.formats.printing", # TODO
+  "pandas.io.formats.style", # TODO
+  "pandas.io.formats.style_render", # TODO
+  "pandas.io.formats.xml", # TODO
+  "pandas.io.json.*", # TODO
+  "pandas.io.parsers.*", # TODO
+  "pandas.io.sas.sas_xport", # TODO
+  "pandas.io.sas.sas7bdat", # TODO
+  "pandas.io.clipboards", # TODO
+  "pandas.io.common", # TODO
+  "pandas.io.gbq", # TODO
+  "pandas.io.html", # TODO
+  "pandas.io.gbq", # TODO
+  "pandas.io.parquet", # TODO
+  "pandas.io.pytables", # TODO
+  "pandas.io.sql", # TODO
+  "pandas.io.stata", # TODO
+  "pandas.io.xml", # TODO
+  "pandas.plotting.*", # TODO
+  "pandas.tests.*",
+  "pandas.tseries.frequencies", # TODO
+  "pandas.tseries.holiday", # TODO
+  "pandas.util._decorators", # TODO
+  "pandas.util._doctools", # TODO
+  "pandas.util._print_versions", # TODO
+  "pandas.util._test_decorators", # TODO
+  "pandas.util._validators", # TODO
+  "pandas.util", # TODO
+  "pandas._version",
+  "pandas.conftest",
+  "pandas"
+]
+disallow_untyped_calls = false
+disallow_untyped_defs = false
+disallow_incomplete_defs = false
+
+[[tool.mypy.overrides]]
+module = [
+  "pandas.tests.*",
+  "pandas._version",
+  "pandas.io.clipboard",
+]
+check_untyped_defs = false
+
+[[tool.mypy.overrides]]
+module = [
+  "pandas.tests.apply.test_series_apply",
+  "pandas.tests.arithmetic.conftest",
+  "pandas.tests.arrays.sparse.test_combine_concat",
+  "pandas.tests.dtypes.test_common",
+  "pandas.tests.frame.methods.test_to_records",
+  "pandas.tests.groupby.test_rank",
+  "pandas.tests.groupby.transform.test_transform",
+  "pandas.tests.indexes.interval.test_interval",
+  "pandas.tests.indexing.test_categorical",
+  "pandas.tests.io.excel.test_writers",
+  "pandas.tests.reductions.test_reductions",
+  "pandas.tests.test_expressions",
+]
+ignore_errors = true
+
+# To be kept consistent with "Import Formatting" section in contributing.rst
+[tool.isort]
+known_pre_libs = "pandas._config"
+known_pre_core = ["pandas._libs", "pandas._typing", "pandas.util._*", "pandas.compat", "pandas.errors"]
+known_dtypes = "pandas.core.dtypes"
+known_post_core = ["pandas.tseries", "pandas.io", "pandas.plotting"]
+sections = ["FUTURE", "STDLIB", "THIRDPARTY" ,"PRE_LIBS" , "PRE_CORE", "DTYPES", "FIRSTPARTY", "POST_CORE", "LOCALFOLDER"]
+profile = "black"
+combine_as_imports = true
+force_grid_wrap = 2
+force_sort_within_sections = true
+skip_glob = "env"
+skip = "pandas/__init__.py"
+
+[tool.pyright]
+pythonVersion = "3.11"
+typeCheckingMode = "basic"
+useLibraryCodeForTypes = false
+include = ["pandas", "typings"]
+exclude = ["pandas/tests", "pandas/io/clipboard", "pandas/util/version", "pandas/core/_numba/extensions.py"]
+# enable subset of "strict"
+reportDuplicateImport = true
+reportInconsistentConstructor = true
+reportInvalidStubStatement = true
+reportOverlappingOverload = true
+reportPropertyTypeMismatch = true
+reportUntypedClassDecorator = true
+reportUntypedFunctionDecorator = true
+reportUntypedNamedTuple = true
+reportUnusedImport = true
+disableBytesTypePromotions = true
+# disable subset of "basic"
+reportGeneralTypeIssues = false
+reportMissingModuleSource = false
+reportOptionalCall = false
+reportOptionalIterable = false
+reportOptionalMemberAccess = false
+reportOptionalOperand = false
+reportOptionalSubscript = false
+reportPrivateImportUsage = false
+reportUnboundVariable = false
+
+[tool.coverage.run]
+branch = true
+omit = ["pandas/_typing.py", "pandas/_version.py"]
+plugins = ["Cython.Coverage"]
+source = ["pandas"]
+
+[tool.coverage.report]
+ignore_errors = false
+show_missing = true
+omit = ["pandas/_version.py"]
+exclude_lines = [
+  # Have to re-enable the standard pragma
+  "pragma: no cover",
+  # Don't complain about missing debug-only code:s
+  "def __repr__",
+  "if self.debug",
+  # Don't complain if tests don't hit defensive assertion code:
+  "raise AssertionError",
+  "raise NotImplementedError",
+  "AbstractMethodError",
+  # Don't complain if non-runnable code isn't run:
+  "if 0:",
+  "if __name__ == .__main__.:",
+  "if TYPE_CHECKING:",
+]
+
+[tool.coverage.html]
+directory = "coverage_html_report"
+
+[tool.codespell]
+ignore-words-list = "blocs, coo, hist, nd, sav, ser, recuse, nin, timere, expec, expecs"
+ignore-regex = 'https://([\w/\.])+'

lib/python3.12/site-packages/pandas/testing.py

@@ -0,0 +1,18 @@
+"""
+Public testing utility functions.
+"""
+
+
+from pandas._testing import (
+    assert_extension_array_equal,
+    assert_frame_equal,
+    assert_index_equal,
+    assert_series_equal,
+)
+
+__all__ = [
+    "assert_extension_array_equal",
+    "assert_frame_equal",
+    "assert_series_equal",
+    "assert_index_equal",
+]

lib/python3.12/site-packages/pandas/tests/computation/test_compat.py

@@ -0,0 +1,32 @@
+import pytest
+
+from pandas.compat._optional import VERSIONS
+
+import pandas as pd
+from pandas.core.computation import expr
+from pandas.core.computation.engines import ENGINES
+from pandas.util.version import Version
+
+
+def test_compat():
+    # test we have compat with our version of numexpr
+
+    from pandas.core.computation.check import NUMEXPR_INSTALLED
+
+    ne = pytest.importorskip("numexpr")
+
+    ver = ne.__version__
+    if Version(ver) < Version(VERSIONS["numexpr"]):
+        assert not NUMEXPR_INSTALLED
+    else:
+        assert NUMEXPR_INSTALLED
+
+
+@pytest.mark.parametrize("engine", ENGINES)
+@pytest.mark.parametrize("parser", expr.PARSERS)
+def test_invalid_numexpr_version(engine, parser):
+    if engine == "numexpr":
+        pytest.importorskip("numexpr")
+    a, b = 1, 2  # noqa: F841
+    res = pd.eval("a + b", engine=engine, parser=parser)
+    assert res == 3

lib/python3.12/site-packages/pandas/tests/computation/test_eval.py

@@ -0,0 +1,2001 @@
+from __future__ import annotations
+
+from functools import reduce
+from itertools import product
+import operator
+
+import numpy as np
+import pytest
+
+from pandas.compat import PY312
+from pandas.errors import (
+    NumExprClobberingError,
+    PerformanceWarning,
+    UndefinedVariableError,
+)
+import pandas.util._test_decorators as td
+
+from pandas.core.dtypes.common import (
+    is_bool,
+    is_float,
+    is_list_like,
+    is_scalar,
+)
+
+import pandas as pd
+from pandas import (
+    DataFrame,
+    Index,
+    Series,
+    date_range,
+    period_range,
+    timedelta_range,
+)
+import pandas._testing as tm
+from pandas.core.computation import (
+    expr,
+    pytables,
+)
+from pandas.core.computation.engines import ENGINES
+from pandas.core.computation.expr import (
+    BaseExprVisitor,
+    PandasExprVisitor,
+    PythonExprVisitor,
+)
+from pandas.core.computation.expressions import (
+    NUMEXPR_INSTALLED,
+    USE_NUMEXPR,
+)
+from pandas.core.computation.ops import (
+    ARITH_OPS_SYMS,
+    SPECIAL_CASE_ARITH_OPS_SYMS,
+    _binary_math_ops,
+    _binary_ops_dict,
+    _unary_math_ops,
+)
+from pandas.core.computation.scope import DEFAULT_GLOBALS
+
+
+@pytest.fixture(
+    params=(
+        pytest.param(
+            engine,
+            marks=[
+                pytest.mark.skipif(
+                    engine == "numexpr" and not USE_NUMEXPR,
+                    reason=f"numexpr enabled->{USE_NUMEXPR}, "
+                    f"installed->{NUMEXPR_INSTALLED}",
+                ),
+                td.skip_if_no("numexpr"),
+            ],
+        )
+        for engine in ENGINES
+    )
+)
+def engine(request):
+    return request.param
+
+
+@pytest.fixture(params=expr.PARSERS)
+def parser(request):
+    return request.param
+
+
+def _eval_single_bin(lhs, cmp1, rhs, engine):
+    c = _binary_ops_dict[cmp1]
+    if ENGINES[engine].has_neg_frac:
+        try:
+            return c(lhs, rhs)
+        except ValueError as e:
+            if str(e).startswith(
+                "negative number cannot be raised to a fractional power"
+            ):
+                return np.nan
+            raise
+    return c(lhs, rhs)
+
+
+# TODO: using range(5) here is a kludge
+@pytest.fixture(
+    params=list(range(5)),
+    ids=["DataFrame", "Series", "SeriesNaN", "DataFrameNaN", "float"],
+)
+def lhs(request):
+    nan_df1 = DataFrame(np.random.default_rng(2).standard_normal((10, 5)))
+    nan_df1[nan_df1 > 0.5] = np.nan
+
+    opts = (
+        DataFrame(np.random.default_rng(2).standard_normal((10, 5))),
+        Series(np.random.default_rng(2).standard_normal(5)),
+        Series([1, 2, np.nan, np.nan, 5]),
+        nan_df1,
+        np.random.default_rng(2).standard_normal(),
+    )
+    return opts[request.param]
+
+
+rhs = lhs
+midhs = lhs
+
+
+@pytest.fixture
+def idx_func_dict():
+    return {
+        "i": lambda n: Index(np.arange(n), dtype=np.int64),
+        "f": lambda n: Index(np.arange(n), dtype=np.float64),
+        "s": lambda n: Index([f"{i}_{chr(i)}" for i in range(97, 97 + n)]),
+        "dt": lambda n: date_range("2020-01-01", periods=n),
+        "td": lambda n: timedelta_range("1 day", periods=n),
+        "p": lambda n: period_range("2020-01-01", periods=n, freq="D"),
+    }
+
+
+class TestEval:
+    @pytest.mark.parametrize(
+        "cmp1",
+        ["!=", "==", "<=", ">=", "<", ">"],
+        ids=["ne", "eq", "le", "ge", "lt", "gt"],
+    )
+    @pytest.mark.parametrize("cmp2", [">", "<"], ids=["gt", "lt"])
+    @pytest.mark.parametrize("binop", expr.BOOL_OPS_SYMS)
+    def test_complex_cmp_ops(self, cmp1, cmp2, binop, lhs, rhs, engine, parser):
+        if parser == "python" and binop in ["and", "or"]:
+            msg = "'BoolOp' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                ex = f"(lhs {cmp1} rhs) {binop} (lhs {cmp2} rhs)"
+                pd.eval(ex, engine=engine, parser=parser)
+            return
+
+        lhs_new = _eval_single_bin(lhs, cmp1, rhs, engine)
+        rhs_new = _eval_single_bin(lhs, cmp2, rhs, engine)
+        expected = _eval_single_bin(lhs_new, binop, rhs_new, engine)
+
+        ex = f"(lhs {cmp1} rhs) {binop} (lhs {cmp2} rhs)"
+        result = pd.eval(ex, engine=engine, parser=parser)
+        tm.assert_equal(result, expected)
+
+    @pytest.mark.parametrize("cmp_op", expr.CMP_OPS_SYMS)
+    def test_simple_cmp_ops(self, cmp_op, lhs, rhs, engine, parser):
+        lhs = lhs < 0
+        rhs = rhs < 0
+
+        if parser == "python" and cmp_op in ["in", "not in"]:
+            msg = "'(In|NotIn)' nodes are not implemented"
+
+            with pytest.raises(NotImplementedError, match=msg):
+                ex = f"lhs {cmp_op} rhs"
+                pd.eval(ex, engine=engine, parser=parser)
+            return
+
+        ex = f"lhs {cmp_op} rhs"
+        msg = "|".join(
+            [
+                r"only list-like( or dict-like)? objects are allowed to be "
+                r"passed to (DataFrame\.)?isin\(\), you passed a "
+                r"(`|')bool(`|')",
+                "argument of type 'bool' is not iterable",
+            ]
+        )
+        if cmp_op in ("in", "not in") and not is_list_like(rhs):
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(
+                    ex,
+                    engine=engine,
+                    parser=parser,
+                    local_dict={"lhs": lhs, "rhs": rhs},
+                )
+        else:
+            expected = _eval_single_bin(lhs, cmp_op, rhs, engine)
+            result = pd.eval(ex, engine=engine, parser=parser)
+            tm.assert_equal(result, expected)
+
+    @pytest.mark.parametrize("op", expr.CMP_OPS_SYMS)
+    def test_compound_invert_op(self, op, lhs, rhs, request, engine, parser):
+        if parser == "python" and op in ["in", "not in"]:
+            msg = "'(In|NotIn)' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                ex = f"~(lhs {op} rhs)"
+                pd.eval(ex, engine=engine, parser=parser)
+            return
+
+        if (
+            is_float(lhs)
+            and not is_float(rhs)
+            and op in ["in", "not in"]
+            and engine == "python"
+            and parser == "pandas"
+        ):
+            mark = pytest.mark.xfail(
+                reason="Looks like expected is negative, unclear whether "
+                "expected is incorrect or result is incorrect"
+            )
+            request.applymarker(mark)
+        skip_these = ["in", "not in"]
+        ex = f"~(lhs {op} rhs)"
+
+        msg = "|".join(
+            [
+                r"only list-like( or dict-like)? objects are allowed to be "
+                r"passed to (DataFrame\.)?isin\(\), you passed a "
+                r"(`|')float(`|')",
+                "argument of type 'float' is not iterable",
+            ]
+        )
+        if is_scalar(rhs) and op in skip_these:
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(
+                    ex,
+                    engine=engine,
+                    parser=parser,
+                    local_dict={"lhs": lhs, "rhs": rhs},
+                )
+        else:
+            # compound
+            if is_scalar(lhs) and is_scalar(rhs):
+                lhs, rhs = (np.array([x]) for x in (lhs, rhs))
+            expected = _eval_single_bin(lhs, op, rhs, engine)
+            if is_scalar(expected):
+                expected = not expected
+            else:
+                expected = ~expected
+            result = pd.eval(ex, engine=engine, parser=parser)
+            tm.assert_almost_equal(expected, result)
+
+    @pytest.mark.parametrize("cmp1", ["<", ">"])
+    @pytest.mark.parametrize("cmp2", ["<", ">"])
+    def test_chained_cmp_op(self, cmp1, cmp2, lhs, midhs, rhs, engine, parser):
+        mid = midhs
+        if parser == "python":
+            ex1 = f"lhs {cmp1} mid {cmp2} rhs"
+            msg = "'BoolOp' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(ex1, engine=engine, parser=parser)
+            return
+
+        lhs_new = _eval_single_bin(lhs, cmp1, mid, engine)
+        rhs_new = _eval_single_bin(mid, cmp2, rhs, engine)
+
+        if lhs_new is not None and rhs_new is not None:
+            ex1 = f"lhs {cmp1} mid {cmp2} rhs"
+            ex2 = f"lhs {cmp1} mid and mid {cmp2} rhs"
+            ex3 = f"(lhs {cmp1} mid) & (mid {cmp2} rhs)"
+            expected = _eval_single_bin(lhs_new, "&", rhs_new, engine)
+
+            for ex in (ex1, ex2, ex3):
+                result = pd.eval(ex, engine=engine, parser=parser)
+
+                tm.assert_almost_equal(result, expected)
+
+    @pytest.mark.parametrize(
+        "arith1", sorted(set(ARITH_OPS_SYMS).difference(SPECIAL_CASE_ARITH_OPS_SYMS))
+    )
+    def test_binary_arith_ops(self, arith1, lhs, rhs, engine, parser):
+        ex = f"lhs {arith1} rhs"
+        result = pd.eval(ex, engine=engine, parser=parser)
+        expected = _eval_single_bin(lhs, arith1, rhs, engine)
+
+        tm.assert_almost_equal(result, expected)
+        ex = f"lhs {arith1} rhs {arith1} rhs"
+        result = pd.eval(ex, engine=engine, parser=parser)
+        nlhs = _eval_single_bin(lhs, arith1, rhs, engine)
+        try:
+            nlhs, ghs = nlhs.align(rhs)
+        except (ValueError, TypeError, AttributeError):
+            # ValueError: series frame or frame series align
+            # TypeError, AttributeError: series or frame with scalar align
+            return
+        else:
+            if engine == "numexpr":
+                import numexpr as ne
+
+                # direct numpy comparison
+                expected = ne.evaluate(f"nlhs {arith1} ghs")
+                # Update assert statement due to unreliable numerical
+                # precision component (GH37328)
+                # TODO: update testing code so that assert_almost_equal statement
+                #  can be replaced again by the assert_numpy_array_equal statement
+                tm.assert_almost_equal(result.values, expected)
+            else:
+                expected = eval(f"nlhs {arith1} ghs")
+                tm.assert_almost_equal(result, expected)
+
+    # modulus, pow, and floor division require special casing
+
+    def test_modulus(self, lhs, rhs, engine, parser):
+        ex = r"lhs % rhs"
+        result = pd.eval(ex, engine=engine, parser=parser)
+        expected = lhs % rhs
+        tm.assert_almost_equal(result, expected)
+
+        if engine == "numexpr":
+            import numexpr as ne
+
+            expected = ne.evaluate(r"expected % rhs")
+            if isinstance(result, (DataFrame, Series)):
+                tm.assert_almost_equal(result.values, expected)
+            else:
+                tm.assert_almost_equal(result, expected.item())
+        else:
+            expected = _eval_single_bin(expected, "%", rhs, engine)
+            tm.assert_almost_equal(result, expected)
+
+    def test_floor_division(self, lhs, rhs, engine, parser):
+        ex = "lhs // rhs"
+
+        if engine == "python":
+            res = pd.eval(ex, engine=engine, parser=parser)
+            expected = lhs // rhs
+            tm.assert_equal(res, expected)
+        else:
+            msg = (
+                r"unsupported operand type\(s\) for //: 'VariableNode' and "
+                "'VariableNode'"
+            )
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(
+                    ex,
+                    local_dict={"lhs": lhs, "rhs": rhs},
+                    engine=engine,
+                    parser=parser,
+                )
+
+    @td.skip_if_windows
+    def test_pow(self, lhs, rhs, engine, parser):
+        # odd failure on win32 platform, so skip
+        ex = "lhs ** rhs"
+        expected = _eval_single_bin(lhs, "**", rhs, engine)
+        result = pd.eval(ex, engine=engine, parser=parser)
+
+        if (
+            is_scalar(lhs)
+            and is_scalar(rhs)
+            and isinstance(expected, (complex, np.complexfloating))
+            and np.isnan(result)
+        ):
+            msg = "(DataFrame.columns|numpy array) are different"
+            with pytest.raises(AssertionError, match=msg):
+                tm.assert_numpy_array_equal(result, expected)
+        else:
+            tm.assert_almost_equal(result, expected)
+
+            ex = "(lhs ** rhs) ** rhs"
+            result = pd.eval(ex, engine=engine, parser=parser)
+
+            middle = _eval_single_bin(lhs, "**", rhs, engine)
+            expected = _eval_single_bin(middle, "**", rhs, engine)
+            tm.assert_almost_equal(result, expected)
+
+    def test_check_single_invert_op(self, lhs, engine, parser):
+        # simple
+        try:
+            elb = lhs.astype(bool)
+        except AttributeError:
+            elb = np.array([bool(lhs)])
+        expected = ~elb
+        result = pd.eval("~elb", engine=engine, parser=parser)
+        tm.assert_almost_equal(expected, result)
+
+    def test_frame_invert(self, engine, parser):
+        expr = "~lhs"
+
+        # ~ ##
+        # frame
+        # float always raises
+        lhs = DataFrame(np.random.default_rng(2).standard_normal((5, 2)))
+        if engine == "numexpr":
+            msg = "couldn't find matching opcode for 'invert_dd'"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            msg = "ufunc 'invert' not supported for the input types"
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+
+        # int raises on numexpr
+        lhs = DataFrame(np.random.default_rng(2).integers(5, size=(5, 2)))
+        if engine == "numexpr":
+            msg = "couldn't find matching opcode for 'invert"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            expect = ~lhs
+            result = pd.eval(expr, engine=engine, parser=parser)
+            tm.assert_frame_equal(expect, result)
+
+        # bool always works
+        lhs = DataFrame(np.random.default_rng(2).standard_normal((5, 2)) > 0.5)
+        expect = ~lhs
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_frame_equal(expect, result)
+
+        # object raises
+        lhs = DataFrame(
+            {"b": ["a", 1, 2.0], "c": np.random.default_rng(2).standard_normal(3) > 0.5}
+        )
+        if engine == "numexpr":
+            with pytest.raises(ValueError, match="unknown type object"):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            msg = "bad operand type for unary ~: 'str'"
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+
+    def test_series_invert(self, engine, parser):
+        # ~ ####
+        expr = "~lhs"
+
+        # series
+        # float raises
+        lhs = Series(np.random.default_rng(2).standard_normal(5))
+        if engine == "numexpr":
+            msg = "couldn't find matching opcode for 'invert_dd'"
+            with pytest.raises(NotImplementedError, match=msg):
+                result = pd.eval(expr, engine=engine, parser=parser)
+        else:
+            msg = "ufunc 'invert' not supported for the input types"
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+
+        # int raises on numexpr
+        lhs = Series(np.random.default_rng(2).integers(5, size=5))
+        if engine == "numexpr":
+            msg = "couldn't find matching opcode for 'invert"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            expect = ~lhs
+            result = pd.eval(expr, engine=engine, parser=parser)
+            tm.assert_series_equal(expect, result)
+
+        # bool
+        lhs = Series(np.random.default_rng(2).standard_normal(5) > 0.5)
+        expect = ~lhs
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_series_equal(expect, result)
+
+        # float
+        # int
+        # bool
+
+        # object
+        lhs = Series(["a", 1, 2.0])
+        if engine == "numexpr":
+            with pytest.raises(ValueError, match="unknown type object"):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            msg = "bad operand type for unary ~: 'str'"
+            with pytest.raises(TypeError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+
+    def test_frame_negate(self, engine, parser):
+        expr = "-lhs"
+
+        # float
+        lhs = DataFrame(np.random.default_rng(2).standard_normal((5, 2)))
+        expect = -lhs
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_frame_equal(expect, result)
+
+        # int
+        lhs = DataFrame(np.random.default_rng(2).integers(5, size=(5, 2)))
+        expect = -lhs
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_frame_equal(expect, result)
+
+        # bool doesn't work with numexpr but works elsewhere
+        lhs = DataFrame(np.random.default_rng(2).standard_normal((5, 2)) > 0.5)
+        if engine == "numexpr":
+            msg = "couldn't find matching opcode for 'neg_bb'"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            expect = -lhs
+            result = pd.eval(expr, engine=engine, parser=parser)
+            tm.assert_frame_equal(expect, result)
+
+    def test_series_negate(self, engine, parser):
+        expr = "-lhs"
+
+        # float
+        lhs = Series(np.random.default_rng(2).standard_normal(5))
+        expect = -lhs
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_series_equal(expect, result)
+
+        # int
+        lhs = Series(np.random.default_rng(2).integers(5, size=5))
+        expect = -lhs
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_series_equal(expect, result)
+
+        # bool doesn't work with numexpr but works elsewhere
+        lhs = Series(np.random.default_rng(2).standard_normal(5) > 0.5)
+        if engine == "numexpr":
+            msg = "couldn't find matching opcode for 'neg_bb'"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(expr, engine=engine, parser=parser)
+        else:
+            expect = -lhs
+            result = pd.eval(expr, engine=engine, parser=parser)
+            tm.assert_series_equal(expect, result)
+
+    @pytest.mark.parametrize(
+        "lhs",
+        [
+            # Float
+            DataFrame(np.random.default_rng(2).standard_normal((5, 2))),
+            # Int
+            DataFrame(np.random.default_rng(2).integers(5, size=(5, 2))),
+            # bool doesn't work with numexpr but works elsewhere
+            DataFrame(np.random.default_rng(2).standard_normal((5, 2)) > 0.5),
+        ],
+    )
+    def test_frame_pos(self, lhs, engine, parser):
+        expr = "+lhs"
+        expect = lhs
+
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_frame_equal(expect, result)
+
+    @pytest.mark.parametrize(
+        "lhs",
+        [
+            # Float
+            Series(np.random.default_rng(2).standard_normal(5)),
+            # Int
+            Series(np.random.default_rng(2).integers(5, size=5)),
+            # bool doesn't work with numexpr but works elsewhere
+            Series(np.random.default_rng(2).standard_normal(5) > 0.5),
+        ],
+    )
+    def test_series_pos(self, lhs, engine, parser):
+        expr = "+lhs"
+        expect = lhs
+
+        result = pd.eval(expr, engine=engine, parser=parser)
+        tm.assert_series_equal(expect, result)
+
+    def test_scalar_unary(self, engine, parser):
+        msg = "bad operand type for unary ~: 'float'"
+        warn = None
+        if PY312 and not (engine == "numexpr" and parser == "pandas"):
+            warn = DeprecationWarning
+        with pytest.raises(TypeError, match=msg):
+            pd.eval("~1.0", engine=engine, parser=parser)
+
+        assert pd.eval("-1.0", parser=parser, engine=engine) == -1.0
+        assert pd.eval("+1.0", parser=parser, engine=engine) == +1.0
+        assert pd.eval("~1", parser=parser, engine=engine) == ~1
+        assert pd.eval("-1", parser=parser, engine=engine) == -1
+        assert pd.eval("+1", parser=parser, engine=engine) == +1
+        with tm.assert_produces_warning(
+            warn, match="Bitwise inversion", check_stacklevel=False
+        ):
+            assert pd.eval("~True", parser=parser, engine=engine) == ~True
+        with tm.assert_produces_warning(
+            warn, match="Bitwise inversion", check_stacklevel=False
+        ):
+            assert pd.eval("~False", parser=parser, engine=engine) == ~False
+        assert pd.eval("-True", parser=parser, engine=engine) == -True
+        assert pd.eval("-False", parser=parser, engine=engine) == -False
+        assert pd.eval("+True", parser=parser, engine=engine) == +True
+        assert pd.eval("+False", parser=parser, engine=engine) == +False
+
+    def test_unary_in_array(self):
+        # GH 11235
+        # TODO: 2022-01-29: result return list with numexpr 2.7.3 in CI
+        # but cannot reproduce locally
+        result = np.array(
+            pd.eval("[-True, True, +True, -False, False, +False, -37, 37, ~37, +37]"),
+            dtype=np.object_,
+        )
+        expected = np.array(
+            [
+                -True,
+                True,
+                +True,
+                -False,
+                False,
+                +False,
+                -37,
+                37,
+                ~37,
+                +37,
+            ],
+            dtype=np.object_,
+        )
+        tm.assert_numpy_array_equal(result, expected)
+
+    @pytest.mark.parametrize("dtype", [np.float32, np.float64])
+    @pytest.mark.parametrize("expr", ["x < -0.1", "-5 > x"])
+    def test_float_comparison_bin_op(self, dtype, expr):
+        # GH 16363
+        df = DataFrame({"x": np.array([0], dtype=dtype)})
+        res = df.eval(expr)
+        assert res.values == np.array([False])
+
+    def test_unary_in_function(self):
+        # GH 46471
+        df = DataFrame({"x": [0, 1, np.nan]})
+
+        result = df.eval("x.fillna(-1)")
+        expected = df.x.fillna(-1)
+        # column name becomes None if using numexpr
+        # only check names when the engine is not numexpr
+        tm.assert_series_equal(result, expected, check_names=not USE_NUMEXPR)
+
+        result = df.eval("x.shift(1, fill_value=-1)")
+        expected = df.x.shift(1, fill_value=-1)
+        tm.assert_series_equal(result, expected, check_names=not USE_NUMEXPR)
+
+    @pytest.mark.parametrize(
+        "ex",
+        (
+            "1 or 2",
+            "1 and 2",
+            "a and b",
+            "a or b",
+            "1 or 2 and (3 + 2) > 3",
+            "2 * x > 2 or 1 and 2",
+            "2 * df > 3 and 1 or a",
+        ),
+    )
+    def test_disallow_scalar_bool_ops(self, ex, engine, parser):
+        x, a, b = np.random.default_rng(2).standard_normal(3), 1, 2  # noqa: F841
+        df = DataFrame(np.random.default_rng(2).standard_normal((3, 2)))  # noqa: F841
+
+        msg = "cannot evaluate scalar only bool ops|'BoolOp' nodes are not"
+        with pytest.raises(NotImplementedError, match=msg):
+            pd.eval(ex, engine=engine, parser=parser)
+
+    def test_identical(self, engine, parser):
+        # see gh-10546
+        x = 1
+        result = pd.eval("x", engine=engine, parser=parser)
+        assert result == 1
+        assert is_scalar(result)
+
+        x = 1.5
+        result = pd.eval("x", engine=engine, parser=parser)
+        assert result == 1.5
+        assert is_scalar(result)
+
+        x = False
+        result = pd.eval("x", engine=engine, parser=parser)
+        assert not result
+        assert is_bool(result)
+        assert is_scalar(result)
+
+        x = np.array([1])
+        result = pd.eval("x", engine=engine, parser=parser)
+        tm.assert_numpy_array_equal(result, np.array([1]))
+        assert result.shape == (1,)
+
+        x = np.array([1.5])
+        result = pd.eval("x", engine=engine, parser=parser)
+        tm.assert_numpy_array_equal(result, np.array([1.5]))
+        assert result.shape == (1,)
+
+        x = np.array([False])  # noqa: F841
+        result = pd.eval("x", engine=engine, parser=parser)
+        tm.assert_numpy_array_equal(result, np.array([False]))
+        assert result.shape == (1,)
+
+    def test_line_continuation(self, engine, parser):
+        # GH 11149
+        exp = """1 + 2 * \
+        5 - 1 + 2 """
+        result = pd.eval(exp, engine=engine, parser=parser)
+        assert result == 12
+
+    def test_float_truncation(self, engine, parser):
+        # GH 14241
+        exp = "1000000000.006"
+        result = pd.eval(exp, engine=engine, parser=parser)
+        expected = np.float64(exp)
+        assert result == expected
+
+        df = DataFrame({"A": [1000000000.0009, 1000000000.0011, 1000000000.0015]})
+        cutoff = 1000000000.0006
+        result = df.query(f"A < {cutoff:.4f}")
+        assert result.empty
+
+        cutoff = 1000000000.0010
+        result = df.query(f"A > {cutoff:.4f}")
+        expected = df.loc[[1, 2], :]
+        tm.assert_frame_equal(expected, result)
+
+        exact = 1000000000.0011
+        result = df.query(f"A == {exact:.4f}")
+        expected = df.loc[[1], :]
+        tm.assert_frame_equal(expected, result)
+
+    def test_disallow_python_keywords(self):
+        # GH 18221
+        df = DataFrame([[0, 0, 0]], columns=["foo", "bar", "class"])
+        msg = "Python keyword not valid identifier in numexpr query"
+        with pytest.raises(SyntaxError, match=msg):
+            df.query("class == 0")
+
+        df = DataFrame()
+        df.index.name = "lambda"
+        with pytest.raises(SyntaxError, match=msg):
+            df.query("lambda == 0")
+
+    def test_true_false_logic(self):
+        # GH 25823
+        # This behavior is deprecated in Python 3.12
+        with tm.maybe_produces_warning(
+            DeprecationWarning, PY312, check_stacklevel=False
+        ):
+            assert pd.eval("not True") == -2
+            assert pd.eval("not False") == -1
+            assert pd.eval("True and not True") == 0
+
+    def test_and_logic_string_match(self):
+        # GH 25823
+        event = Series({"a": "hello"})
+        assert pd.eval(f"{event.str.match('hello').a}")
+        assert pd.eval(f"{event.str.match('hello').a and event.str.match('hello').a}")
+
+
+# -------------------------------------
+# gh-12388: Typecasting rules consistency with python
+
+
+class TestTypeCasting:
+    @pytest.mark.parametrize("op", ["+", "-", "*", "**", "/"])
+    # maybe someday... numexpr has too many upcasting rules now
+    # chain(*(np.core.sctypes[x] for x in ['uint', 'int', 'float']))
+    @pytest.mark.parametrize("left_right", [("df", "3"), ("3", "df")])
+    def test_binop_typecasting(
+        self, engine, parser, op, complex_or_float_dtype, left_right, request
+    ):
+        # GH#21374
+        dtype = complex_or_float_dtype
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)), dtype=dtype)
+        left, right = left_right
+        s = f"{left} {op} {right}"
+        res = pd.eval(s, engine=engine, parser=parser)
+        if dtype == "complex64" and engine == "numexpr":
+            mark = pytest.mark.xfail(
+                reason="numexpr issue with complex that are upcast "
+                "to complex 128 "
+                "https://github.com/pydata/numexpr/issues/492"
+            )
+            request.applymarker(mark)
+        assert df.values.dtype == dtype
+        assert res.values.dtype == dtype
+        tm.assert_frame_equal(res, eval(s), check_exact=False)
+
+
+# -------------------------------------
+# Basic and complex alignment
+
+
+def should_warn(*args):
+    not_mono = not any(map(operator.attrgetter("is_monotonic_increasing"), args))
+    only_one_dt = reduce(
+        operator.xor, (issubclass(x.dtype.type, np.datetime64) for x in args)
+    )
+    return not_mono and only_one_dt
+
+
+class TestAlignment:
+    index_types = ["i", "s", "dt"]
+    lhs_index_types = index_types + ["s"]  # 'p'
+
+    def test_align_nested_unary_op(self, engine, parser):
+        s = "df * ~2"
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))
+        res = pd.eval(s, engine=engine, parser=parser)
+        tm.assert_frame_equal(res, df * ~2)
+
+    @pytest.mark.filterwarnings("always::RuntimeWarning")
+    @pytest.mark.parametrize("lr_idx_type", lhs_index_types)
+    @pytest.mark.parametrize("rr_idx_type", index_types)
+    @pytest.mark.parametrize("c_idx_type", index_types)
+    def test_basic_frame_alignment(
+        self, engine, parser, lr_idx_type, rr_idx_type, c_idx_type, idx_func_dict
+    ):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((10, 10)),
+            index=idx_func_dict[lr_idx_type](10),
+            columns=idx_func_dict[c_idx_type](10),
+        )
+        df2 = DataFrame(
+            np.random.default_rng(2).standard_normal((20, 10)),
+            index=idx_func_dict[rr_idx_type](20),
+            columns=idx_func_dict[c_idx_type](10),
+        )
+        # only warns if not monotonic and not sortable
+        if should_warn(df.index, df2.index):
+            with tm.assert_produces_warning(RuntimeWarning):
+                res = pd.eval("df + df2", engine=engine, parser=parser)
+        else:
+            res = pd.eval("df + df2", engine=engine, parser=parser)
+        tm.assert_frame_equal(res, df + df2)
+
+    @pytest.mark.parametrize("r_idx_type", lhs_index_types)
+    @pytest.mark.parametrize("c_idx_type", lhs_index_types)
+    def test_frame_comparison(
+        self, engine, parser, r_idx_type, c_idx_type, idx_func_dict
+    ):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((10, 10)),
+            index=idx_func_dict[r_idx_type](10),
+            columns=idx_func_dict[c_idx_type](10),
+        )
+        res = pd.eval("df < 2", engine=engine, parser=parser)
+        tm.assert_frame_equal(res, df < 2)
+
+        df3 = DataFrame(
+            np.random.default_rng(2).standard_normal(df.shape),
+            index=df.index,
+            columns=df.columns,
+        )
+        res = pd.eval("df < df3", engine=engine, parser=parser)
+        tm.assert_frame_equal(res, df < df3)
+
+    @pytest.mark.filterwarnings("ignore::RuntimeWarning")
+    @pytest.mark.parametrize("r1", lhs_index_types)
+    @pytest.mark.parametrize("c1", index_types)
+    @pytest.mark.parametrize("r2", index_types)
+    @pytest.mark.parametrize("c2", index_types)
+    def test_medium_complex_frame_alignment(
+        self, engine, parser, r1, c1, r2, c2, idx_func_dict
+    ):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((3, 2)),
+            index=idx_func_dict[r1](3),
+            columns=idx_func_dict[c1](2),
+        )
+        df2 = DataFrame(
+            np.random.default_rng(2).standard_normal((4, 2)),
+            index=idx_func_dict[r2](4),
+            columns=idx_func_dict[c2](2),
+        )
+        df3 = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)),
+            index=idx_func_dict[r2](5),
+            columns=idx_func_dict[c2](2),
+        )
+        if should_warn(df.index, df2.index, df3.index):
+            with tm.assert_produces_warning(RuntimeWarning):
+                res = pd.eval("df + df2 + df3", engine=engine, parser=parser)
+        else:
+            res = pd.eval("df + df2 + df3", engine=engine, parser=parser)
+        tm.assert_frame_equal(res, df + df2 + df3)
+
+    @pytest.mark.filterwarnings("ignore::RuntimeWarning")
+    @pytest.mark.parametrize("index_name", ["index", "columns"])
+    @pytest.mark.parametrize("c_idx_type", index_types)
+    @pytest.mark.parametrize("r_idx_type", lhs_index_types)
+    def test_basic_frame_series_alignment(
+        self, engine, parser, index_name, r_idx_type, c_idx_type, idx_func_dict
+    ):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((10, 10)),
+            index=idx_func_dict[r_idx_type](10),
+            columns=idx_func_dict[c_idx_type](10),
+        )
+        index = getattr(df, index_name)
+        s = Series(np.random.default_rng(2).standard_normal(5), index[:5])
+
+        if should_warn(df.index, s.index):
+            with tm.assert_produces_warning(RuntimeWarning):
+                res = pd.eval("df + s", engine=engine, parser=parser)
+        else:
+            res = pd.eval("df + s", engine=engine, parser=parser)
+
+        if r_idx_type == "dt" or c_idx_type == "dt":
+            expected = df.add(s) if engine == "numexpr" else df + s
+        else:
+            expected = df + s
+        tm.assert_frame_equal(res, expected)
+
+    @pytest.mark.parametrize("index_name", ["index", "columns"])
+    @pytest.mark.parametrize(
+        "r_idx_type, c_idx_type",
+        list(product(["i", "s"], ["i", "s"])) + [("dt", "dt")],
+    )
+    @pytest.mark.filterwarnings("ignore::RuntimeWarning")
+    def test_basic_series_frame_alignment(
+        self, request, engine, parser, index_name, r_idx_type, c_idx_type, idx_func_dict
+    ):
+        if (
+            engine == "numexpr"
+            and parser in ("pandas", "python")
+            and index_name == "index"
+            and r_idx_type == "i"
+            and c_idx_type == "s"
+        ):
+            reason = (
+                f"Flaky column ordering when engine={engine}, "
+                f"parser={parser}, index_name={index_name}, "
+                f"r_idx_type={r_idx_type}, c_idx_type={c_idx_type}"
+            )
+            request.applymarker(pytest.mark.xfail(reason=reason, strict=False))
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((10, 7)),
+            index=idx_func_dict[r_idx_type](10),
+            columns=idx_func_dict[c_idx_type](7),
+        )
+        index = getattr(df, index_name)
+        s = Series(np.random.default_rng(2).standard_normal(5), index[:5])
+        if should_warn(s.index, df.index):
+            with tm.assert_produces_warning(RuntimeWarning):
+                res = pd.eval("s + df", engine=engine, parser=parser)
+        else:
+            res = pd.eval("s + df", engine=engine, parser=parser)
+
+        if r_idx_type == "dt" or c_idx_type == "dt":
+            expected = df.add(s) if engine == "numexpr" else s + df
+        else:
+            expected = s + df
+        tm.assert_frame_equal(res, expected)
+
+    @pytest.mark.filterwarnings("ignore::RuntimeWarning")
+    @pytest.mark.parametrize("c_idx_type", index_types)
+    @pytest.mark.parametrize("r_idx_type", lhs_index_types)
+    @pytest.mark.parametrize("index_name", ["index", "columns"])
+    @pytest.mark.parametrize("op", ["+", "*"])
+    def test_series_frame_commutativity(
+        self, engine, parser, index_name, op, r_idx_type, c_idx_type, idx_func_dict
+    ):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((10, 10)),
+            index=idx_func_dict[r_idx_type](10),
+            columns=idx_func_dict[c_idx_type](10),
+        )
+        index = getattr(df, index_name)
+        s = Series(np.random.default_rng(2).standard_normal(5), index[:5])
+
+        lhs = f"s {op} df"
+        rhs = f"df {op} s"
+        if should_warn(df.index, s.index):
+            with tm.assert_produces_warning(RuntimeWarning):
+                a = pd.eval(lhs, engine=engine, parser=parser)
+            with tm.assert_produces_warning(RuntimeWarning):
+                b = pd.eval(rhs, engine=engine, parser=parser)
+        else:
+            a = pd.eval(lhs, engine=engine, parser=parser)
+            b = pd.eval(rhs, engine=engine, parser=parser)
+
+        if r_idx_type != "dt" and c_idx_type != "dt":
+            if engine == "numexpr":
+                tm.assert_frame_equal(a, b)
+
+    @pytest.mark.filterwarnings("always::RuntimeWarning")
+    @pytest.mark.parametrize("r1", lhs_index_types)
+    @pytest.mark.parametrize("c1", index_types)
+    @pytest.mark.parametrize("r2", index_types)
+    @pytest.mark.parametrize("c2", index_types)
+    def test_complex_series_frame_alignment(
+        self, engine, parser, r1, c1, r2, c2, idx_func_dict
+    ):
+        n = 3
+        m1 = 5
+        m2 = 2 * m1
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((m1, n)),
+            index=idx_func_dict[r1](m1),
+            columns=idx_func_dict[c1](n),
+        )
+        df2 = DataFrame(
+            np.random.default_rng(2).standard_normal((m2, n)),
+            index=idx_func_dict[r2](m2),
+            columns=idx_func_dict[c2](n),
+        )
+        index = df2.columns
+        ser = Series(np.random.default_rng(2).standard_normal(n), index[:n])
+
+        if r2 == "dt" or c2 == "dt":
+            if engine == "numexpr":
+                expected2 = df2.add(ser)
+            else:
+                expected2 = df2 + ser
+        else:
+            expected2 = df2 + ser
+
+        if r1 == "dt" or c1 == "dt":
+            if engine == "numexpr":
+                expected = expected2.add(df)
+            else:
+                expected = expected2 + df
+        else:
+            expected = expected2 + df
+
+        if should_warn(df2.index, ser.index, df.index):
+            with tm.assert_produces_warning(RuntimeWarning):
+                res = pd.eval("df2 + ser + df", engine=engine, parser=parser)
+        else:
+            res = pd.eval("df2 + ser + df", engine=engine, parser=parser)
+        assert res.shape == expected.shape
+        tm.assert_frame_equal(res, expected)
+
+    def test_performance_warning_for_poor_alignment(self, engine, parser):
+        df = DataFrame(np.random.default_rng(2).standard_normal((1000, 10)))
+        s = Series(np.random.default_rng(2).standard_normal(10000))
+        if engine == "numexpr":
+            seen = PerformanceWarning
+        else:
+            seen = False
+
+        with tm.assert_produces_warning(seen):
+            pd.eval("df + s", engine=engine, parser=parser)
+
+        s = Series(np.random.default_rng(2).standard_normal(1000))
+        with tm.assert_produces_warning(False):
+            pd.eval("df + s", engine=engine, parser=parser)
+
+        df = DataFrame(np.random.default_rng(2).standard_normal((10, 10000)))
+        s = Series(np.random.default_rng(2).standard_normal(10000))
+        with tm.assert_produces_warning(False):
+            pd.eval("df + s", engine=engine, parser=parser)
+
+        df = DataFrame(np.random.default_rng(2).standard_normal((10, 10)))
+        s = Series(np.random.default_rng(2).standard_normal(10000))
+
+        is_python_engine = engine == "python"
+
+        if not is_python_engine:
+            wrn = PerformanceWarning
+        else:
+            wrn = False
+
+        with tm.assert_produces_warning(wrn) as w:
+            pd.eval("df + s", engine=engine, parser=parser)
+
+            if not is_python_engine:
+                assert len(w) == 1
+                msg = str(w[0].message)
+                logged = np.log10(s.size - df.shape[1])
+                expected = (
+                    f"Alignment difference on axis 1 is larger "
+                    f"than an order of magnitude on term 'df', "
+                    f"by more than {logged:.4g}; performance may suffer."
+                )
+                assert msg == expected
+
+
+# ------------------------------------
+# Slightly more complex ops
+
+
+class TestOperations:
+    def eval(self, *args, **kwargs):
+        kwargs["level"] = kwargs.pop("level", 0) + 1
+        return pd.eval(*args, **kwargs)
+
+    def test_simple_arith_ops(self, engine, parser):
+        exclude_arith = []
+        if parser == "python":
+            exclude_arith = ["in", "not in"]
+
+        arith_ops = [
+            op
+            for op in expr.ARITH_OPS_SYMS + expr.CMP_OPS_SYMS
+            if op not in exclude_arith
+        ]
+
+        ops = (op for op in arith_ops if op != "//")
+
+        for op in ops:
+            ex = f"1 {op} 1"
+            ex2 = f"x {op} 1"
+            ex3 = f"1 {op} (x + 1)"
+
+            if op in ("in", "not in"):
+                msg = "argument of type 'int' is not iterable"
+                with pytest.raises(TypeError, match=msg):
+                    pd.eval(ex, engine=engine, parser=parser)
+            else:
+                expec = _eval_single_bin(1, op, 1, engine)
+                x = self.eval(ex, engine=engine, parser=parser)
+                assert x == expec
+
+                expec = _eval_single_bin(x, op, 1, engine)
+                y = self.eval(ex2, local_dict={"x": x}, engine=engine, parser=parser)
+                assert y == expec
+
+                expec = _eval_single_bin(1, op, x + 1, engine)
+                y = self.eval(ex3, local_dict={"x": x}, engine=engine, parser=parser)
+                assert y == expec
+
+    @pytest.mark.parametrize("rhs", [True, False])
+    @pytest.mark.parametrize("lhs", [True, False])
+    @pytest.mark.parametrize("op", expr.BOOL_OPS_SYMS)
+    def test_simple_bool_ops(self, rhs, lhs, op):
+        ex = f"{lhs} {op} {rhs}"
+
+        if parser == "python" and op in ["and", "or"]:
+            msg = "'BoolOp' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                self.eval(ex)
+            return
+
+        res = self.eval(ex)
+        exp = eval(ex)
+        assert res == exp
+
+    @pytest.mark.parametrize("rhs", [True, False])
+    @pytest.mark.parametrize("lhs", [True, False])
+    @pytest.mark.parametrize("op", expr.BOOL_OPS_SYMS)
+    def test_bool_ops_with_constants(self, rhs, lhs, op):
+        ex = f"{lhs} {op} {rhs}"
+
+        if parser == "python" and op in ["and", "or"]:
+            msg = "'BoolOp' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                self.eval(ex)
+            return
+
+        res = self.eval(ex)
+        exp = eval(ex)
+        assert res == exp
+
+    def test_4d_ndarray_fails(self):
+        x = np.random.default_rng(2).standard_normal((3, 4, 5, 6))
+        y = Series(np.random.default_rng(2).standard_normal(10))
+        msg = "N-dimensional objects, where N > 2, are not supported with eval"
+        with pytest.raises(NotImplementedError, match=msg):
+            self.eval("x + y", local_dict={"x": x, "y": y})
+
+    def test_constant(self):
+        x = self.eval("1")
+        assert x == 1
+
+    def test_single_variable(self):
+        df = DataFrame(np.random.default_rng(2).standard_normal((10, 2)))
+        df2 = self.eval("df", local_dict={"df": df})
+        tm.assert_frame_equal(df, df2)
+
+    def test_failing_subscript_with_name_error(self):
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))  # noqa: F841
+        with pytest.raises(NameError, match="name 'x' is not defined"):
+            self.eval("df[x > 2] > 2")
+
+    def test_lhs_expression_subscript(self):
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))
+        result = self.eval("(df + 1)[df > 2]", local_dict={"df": df})
+        expected = (df + 1)[df > 2]
+        tm.assert_frame_equal(result, expected)
+
+    def test_attr_expression(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 3)), columns=list("abc")
+        )
+        expr1 = "df.a < df.b"
+        expec1 = df.a < df.b
+        expr2 = "df.a + df.b + df.c"
+        expec2 = df.a + df.b + df.c
+        expr3 = "df.a + df.b + df.c[df.b < 0]"
+        expec3 = df.a + df.b + df.c[df.b < 0]
+        exprs = expr1, expr2, expr3
+        expecs = expec1, expec2, expec3
+        for e, expec in zip(exprs, expecs):
+            tm.assert_series_equal(expec, self.eval(e, local_dict={"df": df}))
+
+    def test_assignment_fails(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 3)), columns=list("abc")
+        )
+        df2 = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))
+        expr1 = "df = df2"
+        msg = "cannot assign without a target object"
+        with pytest.raises(ValueError, match=msg):
+            self.eval(expr1, local_dict={"df": df, "df2": df2})
+
+    def test_assignment_column_multiple_raise(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        # multiple assignees
+        with pytest.raises(SyntaxError, match="invalid syntax"):
+            df.eval("d c = a + b")
+
+    def test_assignment_column_invalid_assign(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        # invalid assignees
+        msg = "left hand side of an assignment must be a single name"
+        with pytest.raises(SyntaxError, match=msg):
+            df.eval("d,c = a + b")
+
+    def test_assignment_column_invalid_assign_function_call(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        msg = "cannot assign to function call"
+        with pytest.raises(SyntaxError, match=msg):
+            df.eval('Timestamp("20131001") = a + b')
+
+    def test_assignment_single_assign_existing(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        # single assignment - existing variable
+        expected = df.copy()
+        expected["a"] = expected["a"] + expected["b"]
+        df.eval("a = a + b", inplace=True)
+        tm.assert_frame_equal(df, expected)
+
+    def test_assignment_single_assign_new(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        # single assignment - new variable
+        expected = df.copy()
+        expected["c"] = expected["a"] + expected["b"]
+        df.eval("c = a + b", inplace=True)
+        tm.assert_frame_equal(df, expected)
+
+    def test_assignment_single_assign_local_overlap(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        df = df.copy()
+        a = 1  # noqa: F841
+        df.eval("a = 1 + b", inplace=True)
+
+        expected = df.copy()
+        expected["a"] = 1 + expected["b"]
+        tm.assert_frame_equal(df, expected)
+
+    def test_assignment_single_assign_name(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+
+        a = 1  # noqa: F841
+        old_a = df.a.copy()
+        df.eval("a = a + b", inplace=True)
+        result = old_a + df.b
+        tm.assert_series_equal(result, df.a, check_names=False)
+        assert result.name is None
+
+    def test_assignment_multiple_raises(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        # multiple assignment
+        df.eval("c = a + b", inplace=True)
+        msg = "can only assign a single expression"
+        with pytest.raises(SyntaxError, match=msg):
+            df.eval("c = a = b")
+
+    def test_assignment_explicit(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+        # explicit targets
+        self.eval("c = df.a + df.b", local_dict={"df": df}, target=df, inplace=True)
+        expected = df.copy()
+        expected["c"] = expected["a"] + expected["b"]
+        tm.assert_frame_equal(df, expected)
+
+    def test_column_in(self):
+        # GH 11235
+        df = DataFrame({"a": [11], "b": [-32]})
+        result = df.eval("a in [11, -32]")
+        expected = Series([True])
+        # TODO: 2022-01-29: Name check failed with numexpr 2.7.3 in CI
+        # but cannot reproduce locally
+        tm.assert_series_equal(result, expected, check_names=False)
+
+    @pytest.mark.xfail(reason="Unknown: Omitted test_ in name prior.")
+    def test_assignment_not_inplace(self):
+        # see gh-9297
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((5, 2)), columns=list("ab")
+        )
+
+        actual = df.eval("c = a + b", inplace=False)
+        assert actual is not None
+
+        expected = df.copy()
+        expected["c"] = expected["a"] + expected["b"]
+        tm.assert_frame_equal(df, expected)
+
+    def test_multi_line_expression(self, warn_copy_on_write):
+        # GH 11149
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+        expected = df.copy()
+
+        expected["c"] = expected["a"] + expected["b"]
+        expected["d"] = expected["c"] + expected["b"]
+        answer = df.eval(
+            """
+        c = a + b
+        d = c + b""",
+            inplace=True,
+        )
+        tm.assert_frame_equal(expected, df)
+        assert answer is None
+
+        expected["a"] = expected["a"] - 1
+        expected["e"] = expected["a"] + 2
+        answer = df.eval(
+            """
+        a = a - 1
+        e = a + 2""",
+            inplace=True,
+        )
+        tm.assert_frame_equal(expected, df)
+        assert answer is None
+
+        # multi-line not valid if not all assignments
+        msg = "Multi-line expressions are only valid if all expressions contain"
+        with pytest.raises(ValueError, match=msg):
+            df.eval(
+                """
+            a = b + 2
+            b - 2""",
+                inplace=False,
+            )
+
+    def test_multi_line_expression_not_inplace(self):
+        # GH 11149
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+        expected = df.copy()
+
+        expected["c"] = expected["a"] + expected["b"]
+        expected["d"] = expected["c"] + expected["b"]
+        df = df.eval(
+            """
+        c = a + b
+        d = c + b""",
+            inplace=False,
+        )
+        tm.assert_frame_equal(expected, df)
+
+        expected["a"] = expected["a"] - 1
+        expected["e"] = expected["a"] + 2
+        df = df.eval(
+            """
+        a = a - 1
+        e = a + 2""",
+            inplace=False,
+        )
+        tm.assert_frame_equal(expected, df)
+
+    def test_multi_line_expression_local_variable(self):
+        # GH 15342
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+        expected = df.copy()
+
+        local_var = 7
+        expected["c"] = expected["a"] * local_var
+        expected["d"] = expected["c"] + local_var
+        answer = df.eval(
+            """
+        c = a * @local_var
+        d = c + @local_var
+        """,
+            inplace=True,
+        )
+        tm.assert_frame_equal(expected, df)
+        assert answer is None
+
+    def test_multi_line_expression_callable_local_variable(self):
+        # 26426
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+
+        def local_func(a, b):
+            return b
+
+        expected = df.copy()
+        expected["c"] = expected["a"] * local_func(1, 7)
+        expected["d"] = expected["c"] + local_func(1, 7)
+        answer = df.eval(
+            """
+        c = a * @local_func(1, 7)
+        d = c + @local_func(1, 7)
+        """,
+            inplace=True,
+        )
+        tm.assert_frame_equal(expected, df)
+        assert answer is None
+
+    def test_multi_line_expression_callable_local_variable_with_kwargs(self):
+        # 26426
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+
+        def local_func(a, b):
+            return b
+
+        expected = df.copy()
+        expected["c"] = expected["a"] * local_func(b=7, a=1)
+        expected["d"] = expected["c"] + local_func(b=7, a=1)
+        answer = df.eval(
+            """
+        c = a * @local_func(b=7, a=1)
+        d = c + @local_func(b=7, a=1)
+        """,
+            inplace=True,
+        )
+        tm.assert_frame_equal(expected, df)
+        assert answer is None
+
+    def test_assignment_in_query(self):
+        # GH 8664
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+        df_orig = df.copy()
+        msg = "cannot assign without a target object"
+        with pytest.raises(ValueError, match=msg):
+            df.query("a = 1")
+        tm.assert_frame_equal(df, df_orig)
+
+    def test_query_inplace(self):
+        # see gh-11149
+        df = DataFrame({"a": [1, 2, 3], "b": [4, 5, 6]})
+        expected = df.copy()
+        expected = expected[expected["a"] == 2]
+        df.query("a == 2", inplace=True)
+        tm.assert_frame_equal(expected, df)
+
+        df = {}
+        expected = {"a": 3}
+
+        self.eval("a = 1 + 2", target=df, inplace=True)
+        tm.assert_dict_equal(df, expected)
+
+    @pytest.mark.parametrize("invalid_target", [1, "cat", [1, 2], np.array([]), (1, 3)])
+    def test_cannot_item_assign(self, invalid_target):
+        msg = "Cannot assign expression output to target"
+        expression = "a = 1 + 2"
+
+        with pytest.raises(ValueError, match=msg):
+            self.eval(expression, target=invalid_target, inplace=True)
+
+        if hasattr(invalid_target, "copy"):
+            with pytest.raises(ValueError, match=msg):
+                self.eval(expression, target=invalid_target, inplace=False)
+
+    @pytest.mark.parametrize("invalid_target", [1, "cat", (1, 3)])
+    def test_cannot_copy_item(self, invalid_target):
+        msg = "Cannot return a copy of the target"
+        expression = "a = 1 + 2"
+
+        with pytest.raises(ValueError, match=msg):
+            self.eval(expression, target=invalid_target, inplace=False)
+
+    @pytest.mark.parametrize("target", [1, "cat", [1, 2], np.array([]), (1, 3), {1: 2}])
+    def test_inplace_no_assignment(self, target):
+        expression = "1 + 2"
+
+        assert self.eval(expression, target=target, inplace=False) == 3
+
+        msg = "Cannot operate inplace if there is no assignment"
+        with pytest.raises(ValueError, match=msg):
+            self.eval(expression, target=target, inplace=True)
+
+    def test_basic_period_index_boolean_expression(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((2, 2)),
+            columns=period_range("2020-01-01", freq="D", periods=2),
+        )
+        e = df < 2
+        r = self.eval("df < 2", local_dict={"df": df})
+        x = df < 2
+
+        tm.assert_frame_equal(r, e)
+        tm.assert_frame_equal(x, e)
+
+    def test_basic_period_index_subscript_expression(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((2, 2)),
+            columns=period_range("2020-01-01", freq="D", periods=2),
+        )
+        r = self.eval("df[df < 2 + 3]", local_dict={"df": df})
+        e = df[df < 2 + 3]
+        tm.assert_frame_equal(r, e)
+
+    def test_nested_period_index_subscript_expression(self):
+        df = DataFrame(
+            np.random.default_rng(2).standard_normal((2, 2)),
+            columns=period_range("2020-01-01", freq="D", periods=2),
+        )
+        r = self.eval("df[df[df < 2] < 2] + df * 2", local_dict={"df": df})
+        e = df[df[df < 2] < 2] + df * 2
+        tm.assert_frame_equal(r, e)
+
+    def test_date_boolean(self, engine, parser):
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))
+        df["dates1"] = date_range("1/1/2012", periods=5)
+        res = self.eval(
+            "df.dates1 < 20130101",
+            local_dict={"df": df},
+            engine=engine,
+            parser=parser,
+        )
+        expec = df.dates1 < "20130101"
+        tm.assert_series_equal(res, expec, check_names=False)
+
+    def test_simple_in_ops(self, engine, parser):
+        if parser != "python":
+            res = pd.eval("1 in [1, 2]", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("2 in (1, 2)", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("3 in (1, 2)", engine=engine, parser=parser)
+            assert not res
+
+            res = pd.eval("3 not in (1, 2)", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("[3] not in (1, 2)", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("[3] in ([3], 2)", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("[[3]] in [[[3]], 2]", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("(3,) in [(3,), 2]", engine=engine, parser=parser)
+            assert res
+
+            res = pd.eval("(3,) not in [(3,), 2]", engine=engine, parser=parser)
+            assert not res
+
+            res = pd.eval("[(3,)] in [[(3,)], 2]", engine=engine, parser=parser)
+            assert res
+        else:
+            msg = "'In' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval("1 in [1, 2]", engine=engine, parser=parser)
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval("2 in (1, 2)", engine=engine, parser=parser)
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval("3 in (1, 2)", engine=engine, parser=parser)
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval("[(3,)] in (1, 2, [(3,)])", engine=engine, parser=parser)
+            msg = "'NotIn' nodes are not implemented"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval("3 not in (1, 2)", engine=engine, parser=parser)
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval("[3] not in (1, 2, [[3]])", engine=engine, parser=parser)
+
+    def test_check_many_exprs(self, engine, parser):
+        a = 1  # noqa: F841
+        expr = " * ".join("a" * 33)
+        expected = 1
+        res = pd.eval(expr, engine=engine, parser=parser)
+        assert res == expected
+
+    @pytest.mark.parametrize(
+        "expr",
+        [
+            "df > 2 and df > 3",
+            "df > 2 or df > 3",
+            "not df > 2",
+        ],
+    )
+    def test_fails_and_or_not(self, expr, engine, parser):
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))
+        if parser == "python":
+            msg = "'BoolOp' nodes are not implemented"
+            if "not" in expr:
+                msg = "'Not' nodes are not implemented"
+
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(
+                    expr,
+                    local_dict={"df": df},
+                    parser=parser,
+                    engine=engine,
+                )
+        else:
+            # smoke-test, should not raise
+            pd.eval(
+                expr,
+                local_dict={"df": df},
+                parser=parser,
+                engine=engine,
+            )
+
+    @pytest.mark.parametrize("char", ["|", "&"])
+    def test_fails_ampersand_pipe(self, char, engine, parser):
+        df = DataFrame(np.random.default_rng(2).standard_normal((5, 3)))  # noqa: F841
+        ex = f"(df + 2)[df > 1] > 0 {char} (df > 0)"
+        if parser == "python":
+            msg = "cannot evaluate scalar only bool ops"
+            with pytest.raises(NotImplementedError, match=msg):
+                pd.eval(ex, parser=parser, engine=engine)
+        else:
+            # smoke-test, should not raise
+            pd.eval(ex, parser=parser, engine=engine)
+
+
+class TestMath:
+    def eval(self, *args, **kwargs):
+        kwargs["level"] = kwargs.pop("level", 0) + 1
+        return pd.eval(*args, **kwargs)
+
+    @pytest.mark.skipif(
+        not NUMEXPR_INSTALLED, reason="Unary ops only implemented for numexpr"
+    )
+    @pytest.mark.parametrize("fn", _unary_math_ops)
+    def test_unary_functions(self, fn):
+        df = DataFrame({"a": np.random.default_rng(2).standard_normal(10)})
+        a = df.a
+
+        expr = f"{fn}(a)"
+        got = self.eval(expr)
+        with np.errstate(all="ignore"):
+            expect = getattr(np, fn)(a)
+        tm.assert_series_equal(got, expect, check_names=False)
+
+    @pytest.mark.parametrize("fn", _binary_math_ops)
+    def test_binary_functions(self, fn):
+        df = DataFrame(
+            {
+                "a": np.random.default_rng(2).standard_normal(10),
+                "b": np.random.default_rng(2).standard_normal(10),
+            }
+        )
+        a = df.a
+        b = df.b
+
+        expr = f"{fn}(a, b)"
+        got = self.eval(expr)
+        with np.errstate(all="ignore"):
+            expect = getattr(np, fn)(a, b)
+        tm.assert_almost_equal(got, expect, check_names=False)
+
+    def test_df_use_case(self, engine, parser):
+        df = DataFrame(
+            {
+                "a": np.random.default_rng(2).standard_normal(10),
+                "b": np.random.default_rng(2).standard_normal(10),
+            }
+        )
+        df.eval(
+            "e = arctan2(sin(a), b)",
+            engine=engine,
+            parser=parser,
+            inplace=True,
+        )
+        got = df.e
+        expect = np.arctan2(np.sin(df.a), df.b)
+        tm.assert_series_equal(got, expect, check_names=False)
+
+    def test_df_arithmetic_subexpression(self, engine, parser):
+        df = DataFrame(
+            {
+                "a": np.random.default_rng(2).standard_normal(10),
+                "b": np.random.default_rng(2).standard_normal(10),
+            }
+        )
+        df.eval("e = sin(a + b)", engine=engine, parser=parser, inplace=True)
+        got = df.e
+        expect = np.sin(df.a + df.b)
+        tm.assert_series_equal(got, expect, check_names=False)
+
+    @pytest.mark.parametrize(
+        "dtype, expect_dtype",
+        [
+            (np.int32, np.float64),
+            (np.int64, np.float64),
+            (np.float32, np.float32),
+            (np.float64, np.float64),
+            pytest.param(np.complex128, np.complex128, marks=td.skip_if_windows),
+        ],
+    )
+    def test_result_types(self, dtype, expect_dtype, engine, parser):
+        # xref https://github.com/pandas-dev/pandas/issues/12293
+        #  this fails on Windows, apparently a floating point precision issue
+
+        # Did not test complex64 because DataFrame is converting it to
+        # complex128. Due to https://github.com/pandas-dev/pandas/issues/10952
+        df = DataFrame(
+            {"a": np.random.default_rng(2).standard_normal(10).astype(dtype)}
+        )
+        assert df.a.dtype == dtype
+        df.eval("b = sin(a)", engine=engine, parser=parser, inplace=True)
+        got = df.b
+        expect = np.sin(df.a)
+        assert expect.dtype == got.dtype
+        assert expect_dtype == got.dtype
+        tm.assert_series_equal(got, expect, check_names=False)
+
+    def test_undefined_func(self, engine, parser):
+        df = DataFrame({"a": np.random.default_rng(2).standard_normal(10)})
+        msg = '"mysin" is not a supported function'
+
+        with pytest.raises(ValueError, match=msg):
+            df.eval("mysin(a)", engine=engine, parser=parser)
+
+    def test_keyword_arg(self, engine, parser):
+        df = DataFrame({"a": np.random.default_rng(2).standard_normal(10)})
+        msg = 'Function "sin" does not support keyword arguments'
+
+        with pytest.raises(TypeError, match=msg):
+            df.eval("sin(x=a)", engine=engine, parser=parser)
+
+
+_var_s = np.random.default_rng(2).standard_normal(10)
+
+
+class TestScope:
+    def test_global_scope(self, engine, parser):
+        e = "_var_s * 2"
+        tm.assert_numpy_array_equal(
+            _var_s * 2, pd.eval(e, engine=engine, parser=parser)
+        )
+
+    def test_no_new_locals(self, engine, parser):
+        x = 1
+        lcls = locals().copy()
+        pd.eval("x + 1", local_dict=lcls, engine=engine, parser=parser)
+        lcls2 = locals().copy()
+        lcls2.pop("lcls")
+        assert lcls == lcls2
+
+    def test_no_new_globals(self, engine, parser):
+        x = 1  # noqa: F841
+        gbls = globals().copy()
+        pd.eval("x + 1", engine=engine, parser=parser)
+        gbls2 = globals().copy()
+        assert gbls == gbls2
+
+    def test_empty_locals(self, engine, parser):
+        # GH 47084
+        x = 1  # noqa: F841
+        msg = "name 'x' is not defined"
+        with pytest.raises(UndefinedVariableError, match=msg):
+            pd.eval("x + 1", engine=engine, parser=parser, local_dict={})
+
+    def test_empty_globals(self, engine, parser):
+        # GH 47084
+        msg = "name '_var_s' is not defined"
+        e = "_var_s * 2"
+        with pytest.raises(UndefinedVariableError, match=msg):
+            pd.eval(e, engine=engine, parser=parser, global_dict={})
+
+
+@td.skip_if_no("numexpr")
+def test_invalid_engine():
+    msg = "Invalid engine 'asdf' passed"
+    with pytest.raises(KeyError, match=msg):
+        pd.eval("x + y", local_dict={"x": 1, "y": 2}, engine="asdf")
+
+
+@td.skip_if_no("numexpr")
+@pytest.mark.parametrize(
+    ("use_numexpr", "expected"),
+    (
+        (True, "numexpr"),
+        (False, "python"),
+    ),
+)
+def test_numexpr_option_respected(use_numexpr, expected):
+    # GH 32556
+    from pandas.core.computation.eval import _check_engine
+
+    with pd.option_context("compute.use_numexpr", use_numexpr):
+        result = _check_engine(None)
+        assert result == expected
+
+
+@td.skip_if_no("numexpr")
+def test_numexpr_option_incompatible_op():
+    # GH 32556
+    with pd.option_context("compute.use_numexpr", False):
+        df = DataFrame(
+            {"A": [True, False, True, False, None, None], "B": [1, 2, 3, 4, 5, 6]}
+        )
+        result = df.query("A.isnull()")
+        expected = DataFrame({"A": [None, None], "B": [5, 6]}, index=[4, 5])
+        tm.assert_frame_equal(result, expected)
+
+
+@td.skip_if_no("numexpr")
+def test_invalid_parser():
+    msg = "Invalid parser 'asdf' passed"
+    with pytest.raises(KeyError, match=msg):
+        pd.eval("x + y", local_dict={"x": 1, "y": 2}, parser="asdf")
+
+
+_parsers: dict[str, type[BaseExprVisitor]] = {
+    "python": PythonExprVisitor,
+    "pytables": pytables.PyTablesExprVisitor,
+    "pandas": PandasExprVisitor,
+}
+
+
+@pytest.mark.parametrize("engine", ENGINES)
+@pytest.mark.parametrize("parser", _parsers)
+def test_disallowed_nodes(engine, parser):
+    VisitorClass = _parsers[parser]
+    inst = VisitorClass("x + 1", engine, parser)
+
+    for ops in VisitorClass.unsupported_nodes:
+        msg = "nodes are not implemented"
+        with pytest.raises(NotImplementedError, match=msg):
+            getattr(inst, ops)()
+
+
+def test_syntax_error_exprs(engine, parser):
+    e = "s +"
+    with pytest.raises(SyntaxError, match="invalid syntax"):
+        pd.eval(e, engine=engine, parser=parser)
+
+
+def test_name_error_exprs(engine, parser):
+    e = "s + t"
+    msg = "name 's' is not defined"
+    with pytest.raises(NameError, match=msg):
+        pd.eval(e, engine=engine, parser=parser)
+
+
+@pytest.mark.parametrize("express", ["a + @b", "@a + b", "@a + @b"])
+def test_invalid_local_variable_reference(engine, parser, express):
+    a, b = 1, 2  # noqa: F841
+
+    if parser != "pandas":
+        with pytest.raises(SyntaxError, match="The '@' prefix is only"):
+            pd.eval(express, engine=engine, parser=parser)
+    else:
+        with pytest.raises(SyntaxError, match="The '@' prefix is not"):
+            pd.eval(express, engine=engine, parser=parser)
+
+
+def test_numexpr_builtin_raises(engine, parser):
+    sin, dotted_line = 1, 2
+    if engine == "numexpr":
+        msg = "Variables in expression .+"
+        with pytest.raises(NumExprClobberingError, match=msg):
+            pd.eval("sin + dotted_line", engine=engine, parser=parser)
+    else:
+        res = pd.eval("sin + dotted_line", engine=engine, parser=parser)
+        assert res == sin + dotted_line
+
+
+def test_bad_resolver_raises(engine, parser):
+    cannot_resolve = 42, 3.0
+    with pytest.raises(TypeError, match="Resolver of type .+"):
+        pd.eval("1 + 2", resolvers=cannot_resolve, engine=engine, parser=parser)
+
+
+def test_empty_string_raises(engine, parser):
+    # GH 13139
+    with pytest.raises(ValueError, match="expr cannot be an empty string"):
+        pd.eval("", engine=engine, parser=parser)
+
+
+def test_more_than_one_expression_raises(engine, parser):
+    with pytest.raises(SyntaxError, match="only a single expression is allowed"):
+        pd.eval("1 + 1; 2 + 2", engine=engine, parser=parser)
+
+
+@pytest.mark.parametrize("cmp", ("and", "or"))
+@pytest.mark.parametrize("lhs", (int, float))
+@pytest.mark.parametrize("rhs", (int, float))
+def test_bool_ops_fails_on_scalars(lhs, cmp, rhs, engine, parser):
+    gen = {
+        int: lambda: np.random.default_rng(2).integers(10),
+        float: np.random.default_rng(2).standard_normal,
+    }
+
+    mid = gen[lhs]()  # noqa: F841
+    lhs = gen[lhs]()
+    rhs = gen[rhs]()
+
+    ex1 = f"lhs {cmp} mid {cmp} rhs"
+    ex2 = f"lhs {cmp} mid and mid {cmp} rhs"
+    ex3 = f"(lhs {cmp} mid) & (mid {cmp} rhs)"
+    for ex in (ex1, ex2, ex3):
+        msg = "cannot evaluate scalar only bool ops|'BoolOp' nodes are not"
+        with pytest.raises(NotImplementedError, match=msg):
+            pd.eval(ex, engine=engine, parser=parser)
+
+
+@pytest.mark.parametrize(
+    "other",
+    [
+        "'x'",
+        "...",
+    ],
+)
+def test_equals_various(other):
+    df = DataFrame({"A": ["a", "b", "c"]}, dtype=object)
+    result = df.eval(f"A == {other}")
+    expected = Series([False, False, False], name="A")
+    if USE_NUMEXPR:
+        # https://github.com/pandas-dev/pandas/issues/10239
+        # lose name with numexpr engine. Remove when that's fixed.
+        expected.name = None
+    tm.assert_series_equal(result, expected)
+
+
+def test_inf(engine, parser):
+    s = "inf + 1"
+    expected = np.inf
+    result = pd.eval(s, engine=engine, parser=parser)
+    assert result == expected
+
+
+@pytest.mark.parametrize("column", ["Temp(°C)", "Capacitance(μF)"])
+def test_query_token(engine, column):
+    # See: https://github.com/pandas-dev/pandas/pull/42826
+    df = DataFrame(
+        np.random.default_rng(2).standard_normal((5, 2)), columns=[column, "b"]
+    )
+    expected = df[df[column] > 5]
+    query_string = f"`{column}` > 5"
+    result = df.query(query_string, engine=engine)
+    tm.assert_frame_equal(result, expected)
+
+
+def test_negate_lt_eq_le(engine, parser):
+    df = DataFrame([[0, 10], [1, 20]], columns=["cat", "count"])
+    expected = df[~(df.cat > 0)]
+
+    result = df.query("~(cat > 0)", engine=engine, parser=parser)
+    tm.assert_frame_equal(result, expected)
+
+    if parser == "python":
+        msg = "'Not' nodes are not implemented"
+        with pytest.raises(NotImplementedError, match=msg):
+            df.query("not (cat > 0)", engine=engine, parser=parser)
+    else:
+        result = df.query("not (cat > 0)", engine=engine, parser=parser)
+        tm.assert_frame_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    "column",
+    DEFAULT_GLOBALS.keys(),
+)
+def test_eval_no_support_column_name(request, column):
+    # GH 44603
+    if column in ["True", "False", "inf", "Inf"]:
+        request.applymarker(
+            pytest.mark.xfail(
+                raises=KeyError,
+                reason=f"GH 47859 DataFrame eval not supported with {column}",
+            )
+        )
+
+    df = DataFrame(
+        np.random.default_rng(2).integers(0, 100, size=(10, 2)),
+        columns=[column, "col1"],
+    )
+    expected = df[df[column] > 6]
+    result = df.query(f"{column}>6")
+
+    tm.assert_frame_equal(result, expected)
+
+
+def test_set_inplace(using_copy_on_write, warn_copy_on_write):
+    # https://github.com/pandas-dev/pandas/issues/47449
+    # Ensure we don't only update the DataFrame inplace, but also the actual
+    # column values, such that references to this column also get updated
+    df = DataFrame({"A": [1, 2, 3], "B": [4, 5, 6], "C": [7, 8, 9]})
+    result_view = df[:]
+    ser = df["A"]
+    with tm.assert_cow_warning(warn_copy_on_write):
+        df.eval("A = B + C", inplace=True)
+    expected = DataFrame({"A": [11, 13, 15], "B": [4, 5, 6], "C": [7, 8, 9]})
+    tm.assert_frame_equal(df, expected)
+    if not using_copy_on_write:
+        tm.assert_series_equal(ser, expected["A"])
+        tm.assert_series_equal(result_view["A"], expected["A"])
+    else:
+        expected = Series([1, 2, 3], name="A")
+        tm.assert_series_equal(ser, expected)
+        tm.assert_series_equal(result_view["A"], expected)
+
+
+class TestValidate:
+    @pytest.mark.parametrize("value", [1, "True", [1, 2, 3], 5.0])
+    def test_validate_bool_args(self, value):
+        msg = 'For argument "inplace" expected type bool, received type'
+        with pytest.raises(ValueError, match=msg):
+            pd.eval("2+2", inplace=value)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_append.py

@@ -0,0 +1,62 @@
+import pytest
+
+from pandas import (
+    CategoricalIndex,
+    Index,
+)
+import pandas._testing as tm
+
+
+class TestAppend:
+    @pytest.fixture
+    def ci(self):
+        categories = list("cab")
+        return CategoricalIndex(list("aabbca"), categories=categories, ordered=False)
+
+    def test_append(self, ci):
+        # append cats with the same categories
+        result = ci[:3].append(ci[3:])
+        tm.assert_index_equal(result, ci, exact=True)
+
+        foos = [ci[:1], ci[1:3], ci[3:]]
+        result = foos[0].append(foos[1:])
+        tm.assert_index_equal(result, ci, exact=True)
+
+    def test_append_empty(self, ci):
+        # empty
+        result = ci.append([])
+        tm.assert_index_equal(result, ci, exact=True)
+
+    def test_append_mismatched_categories(self, ci):
+        # appending with different categories or reordered is not ok
+        msg = "all inputs must be Index"
+        with pytest.raises(TypeError, match=msg):
+            ci.append(ci.values.set_categories(list("abcd")))
+        with pytest.raises(TypeError, match=msg):
+            ci.append(ci.values.reorder_categories(list("abc")))
+
+    def test_append_category_objects(self, ci):
+        # with objects
+        result = ci.append(Index(["c", "a"]))
+        expected = CategoricalIndex(list("aabbcaca"), categories=ci.categories)
+        tm.assert_index_equal(result, expected, exact=True)
+
+    def test_append_non_categories(self, ci):
+        # invalid objects -> cast to object via concat_compat
+        result = ci.append(Index(["a", "d"]))
+        expected = Index(["a", "a", "b", "b", "c", "a", "a", "d"])
+        tm.assert_index_equal(result, expected, exact=True)
+
+    def test_append_object(self, ci):
+        # GH#14298 - if base object is not categorical -> coerce to object
+        result = Index(["c", "a"]).append(ci)
+        expected = Index(list("caaabbca"))
+        tm.assert_index_equal(result, expected, exact=True)
+
+    def test_append_to_another(self):
+        # hits Index._concat
+        fst = Index(["a", "b"])
+        snd = CategoricalIndex(["d", "e"])
+        result = fst.append(snd)
+        expected = Index(["a", "b", "d", "e"])
+        tm.assert_index_equal(result, expected)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_astype.py

@@ -0,0 +1,90 @@
+from datetime import date
+
+import numpy as np
+import pytest
+
+from pandas import (
+    Categorical,
+    CategoricalDtype,
+    CategoricalIndex,
+    Index,
+    IntervalIndex,
+)
+import pandas._testing as tm
+
+
+class TestAstype:
+    def test_astype(self):
+        ci = CategoricalIndex(list("aabbca"), categories=list("cab"), ordered=False)
+
+        result = ci.astype(object)
+        tm.assert_index_equal(result, Index(np.array(ci), dtype=object))
+
+        # this IS equal, but not the same class
+        assert result.equals(ci)
+        assert isinstance(result, Index)
+        assert not isinstance(result, CategoricalIndex)
+
+        # interval
+        ii = IntervalIndex.from_arrays(left=[-0.001, 2.0], right=[2, 4], closed="right")
+
+        ci = CategoricalIndex(
+            Categorical.from_codes([0, 1, -1], categories=ii, ordered=True)
+        )
+
+        result = ci.astype("interval")
+        expected = ii.take([0, 1, -1], allow_fill=True, fill_value=np.nan)
+        tm.assert_index_equal(result, expected)
+
+        result = IntervalIndex(result.values)
+        tm.assert_index_equal(result, expected)
+
+    @pytest.mark.parametrize("name", [None, "foo"])
+    @pytest.mark.parametrize("dtype_ordered", [True, False])
+    @pytest.mark.parametrize("index_ordered", [True, False])
+    def test_astype_category(self, name, dtype_ordered, index_ordered):
+        # GH#18630
+        index = CategoricalIndex(
+            list("aabbca"), categories=list("cab"), ordered=index_ordered
+        )
+        if name:
+            index = index.rename(name)
+
+        # standard categories
+        dtype = CategoricalDtype(ordered=dtype_ordered)
+        result = index.astype(dtype)
+        expected = CategoricalIndex(
+            index.tolist(),
+            name=name,
+            categories=index.categories,
+            ordered=dtype_ordered,
+        )
+        tm.assert_index_equal(result, expected)
+
+        # non-standard categories
+        dtype = CategoricalDtype(index.unique().tolist()[:-1], dtype_ordered)
+        result = index.astype(dtype)
+        expected = CategoricalIndex(index.tolist(), name=name, dtype=dtype)
+        tm.assert_index_equal(result, expected)
+
+        if dtype_ordered is False:
+            # dtype='category' can't specify ordered, so only test once
+            result = index.astype("category")
+            expected = index
+            tm.assert_index_equal(result, expected)
+
+    @pytest.mark.parametrize("box", [True, False])
+    def test_categorical_date_roundtrip(self, box):
+        # astype to categorical and back should preserve date objects
+        v = date.today()
+
+        obj = Index([v, v])
+        assert obj.dtype == object
+        if box:
+            obj = obj.array
+
+        cat = obj.astype("category")
+
+        rtrip = cat.astype(object)
+        assert rtrip.dtype == object
+        assert type(rtrip[0]) is date

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_category.py

@@ -0,0 +1,394 @@
+import numpy as np
+import pytest
+
+from pandas._config import using_pyarrow_string_dtype
+
+from pandas._libs import index as libindex
+from pandas._libs.arrays import NDArrayBacked
+
+import pandas as pd
+from pandas import (
+    Categorical,
+    CategoricalDtype,
+)
+import pandas._testing as tm
+from pandas.core.indexes.api import (
+    CategoricalIndex,
+    Index,
+)
+
+
+class TestCategoricalIndex:
+    @pytest.fixture
+    def simple_index(self) -> CategoricalIndex:
+        return CategoricalIndex(list("aabbca"), categories=list("cab"), ordered=False)
+
+    def test_can_hold_identifiers(self):
+        idx = CategoricalIndex(list("aabbca"), categories=None, ordered=False)
+        key = idx[0]
+        assert idx._can_hold_identifiers_and_holds_name(key) is True
+
+    def test_insert(self, simple_index):
+        ci = simple_index
+        categories = ci.categories
+
+        # test 0th element
+        result = ci.insert(0, "a")
+        expected = CategoricalIndex(list("aaabbca"), categories=categories)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # test Nth element that follows Python list behavior
+        result = ci.insert(-1, "a")
+        expected = CategoricalIndex(list("aabbcaa"), categories=categories)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # test empty
+        result = CategoricalIndex([], categories=categories).insert(0, "a")
+        expected = CategoricalIndex(["a"], categories=categories)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # invalid -> cast to object
+        expected = ci.astype(object).insert(0, "d")
+        result = ci.insert(0, "d").astype(object)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # GH 18295 (test missing)
+        expected = CategoricalIndex(["a", np.nan, "a", "b", "c", "b"])
+        for na in (np.nan, pd.NaT, None):
+            result = CategoricalIndex(list("aabcb")).insert(1, na)
+            tm.assert_index_equal(result, expected)
+
+    def test_insert_na_mismatched_dtype(self):
+        ci = CategoricalIndex([0, 1, 1])
+        result = ci.insert(0, pd.NaT)
+        expected = Index([pd.NaT, 0, 1, 1], dtype=object)
+        tm.assert_index_equal(result, expected)
+
+    def test_delete(self, simple_index):
+        ci = simple_index
+        categories = ci.categories
+
+        result = ci.delete(0)
+        expected = CategoricalIndex(list("abbca"), categories=categories)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        result = ci.delete(-1)
+        expected = CategoricalIndex(list("aabbc"), categories=categories)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        with tm.external_error_raised((IndexError, ValueError)):
+            # Either depending on NumPy version
+            ci.delete(10)
+
+    @pytest.mark.parametrize(
+        "data, non_lexsorted_data",
+        [[[1, 2, 3], [9, 0, 1, 2, 3]], [list("abc"), list("fabcd")]],
+    )
+    def test_is_monotonic(self, data, non_lexsorted_data):
+        c = CategoricalIndex(data)
+        assert c.is_monotonic_increasing is True
+        assert c.is_monotonic_decreasing is False
+
+        c = CategoricalIndex(data, ordered=True)
+        assert c.is_monotonic_increasing is True
+        assert c.is_monotonic_decreasing is False
+
+        c = CategoricalIndex(data, categories=reversed(data))
+        assert c.is_monotonic_increasing is False
+        assert c.is_monotonic_decreasing is True
+
+        c = CategoricalIndex(data, categories=reversed(data), ordered=True)
+        assert c.is_monotonic_increasing is False
+        assert c.is_monotonic_decreasing is True
+
+        # test when data is neither monotonic increasing nor decreasing
+        reordered_data = [data[0], data[2], data[1]]
+        c = CategoricalIndex(reordered_data, categories=reversed(data))
+        assert c.is_monotonic_increasing is False
+        assert c.is_monotonic_decreasing is False
+
+        # non lexsorted categories
+        categories = non_lexsorted_data
+
+        c = CategoricalIndex(categories[:2], categories=categories)
+        assert c.is_monotonic_increasing is True
+        assert c.is_monotonic_decreasing is False
+
+        c = CategoricalIndex(categories[1:3], categories=categories)
+        assert c.is_monotonic_increasing is True
+        assert c.is_monotonic_decreasing is False
+
+    def test_has_duplicates(self):
+        idx = CategoricalIndex([0, 0, 0], name="foo")
+        assert idx.is_unique is False
+        assert idx.has_duplicates is True
+
+        idx = CategoricalIndex([0, 1], categories=[2, 3], name="foo")
+        assert idx.is_unique is False
+        assert idx.has_duplicates is True
+
+        idx = CategoricalIndex([0, 1, 2, 3], categories=[1, 2, 3], name="foo")
+        assert idx.is_unique is True
+        assert idx.has_duplicates is False
+
+    @pytest.mark.parametrize(
+        "data, categories, expected",
+        [
+            (
+                [1, 1, 1],
+                [1, 2, 3],
+                {
+                    "first": np.array([False, True, True]),
+                    "last": np.array([True, True, False]),
+                    False: np.array([True, True, True]),
+                },
+            ),
+            (
+                [1, 1, 1],
+                list("abc"),
+                {
+                    "first": np.array([False, True, True]),
+                    "last": np.array([True, True, False]),
+                    False: np.array([True, True, True]),
+                },
+            ),
+            (
+                [2, "a", "b"],
+                list("abc"),
+                {
+                    "first": np.zeros(shape=(3), dtype=np.bool_),
+                    "last": np.zeros(shape=(3), dtype=np.bool_),
+                    False: np.zeros(shape=(3), dtype=np.bool_),
+                },
+            ),
+            (
+                list("abb"),
+                list("abc"),
+                {
+                    "first": np.array([False, False, True]),
+                    "last": np.array([False, True, False]),
+                    False: np.array([False, True, True]),
+                },
+            ),
+        ],
+    )
+    def test_drop_duplicates(self, data, categories, expected):
+        idx = CategoricalIndex(data, categories=categories, name="foo")
+        for keep, e in expected.items():
+            tm.assert_numpy_array_equal(idx.duplicated(keep=keep), e)
+            e = idx[~e]
+            result = idx.drop_duplicates(keep=keep)
+            tm.assert_index_equal(result, e)
+
+    @pytest.mark.parametrize(
+        "data, categories, expected_data",
+        [
+            ([1, 1, 1], [1, 2, 3], [1]),
+            ([1, 1, 1], list("abc"), [np.nan]),
+            ([1, 2, "a"], [1, 2, 3], [1, 2, np.nan]),
+            ([2, "a", "b"], list("abc"), [np.nan, "a", "b"]),
+        ],
+    )
+    def test_unique(self, data, categories, expected_data, ordered):
+        dtype = CategoricalDtype(categories, ordered=ordered)
+
+        idx = CategoricalIndex(data, dtype=dtype)
+        expected = CategoricalIndex(expected_data, dtype=dtype)
+        tm.assert_index_equal(idx.unique(), expected)
+
+    @pytest.mark.xfail(using_pyarrow_string_dtype(), reason="repr doesn't roundtrip")
+    def test_repr_roundtrip(self):
+        ci = CategoricalIndex(["a", "b"], categories=["a", "b"], ordered=True)
+        str(ci)
+        tm.assert_index_equal(eval(repr(ci)), ci, exact=True)
+
+        # formatting
+        str(ci)
+
+        # long format
+        # this is not reprable
+        ci = CategoricalIndex(np.random.default_rng(2).integers(0, 5, size=100))
+        str(ci)
+
+    def test_isin(self):
+        ci = CategoricalIndex(list("aabca") + [np.nan], categories=["c", "a", "b"])
+        tm.assert_numpy_array_equal(
+            ci.isin(["c"]), np.array([False, False, False, True, False, False])
+        )
+        tm.assert_numpy_array_equal(
+            ci.isin(["c", "a", "b"]), np.array([True] * 5 + [False])
+        )
+        tm.assert_numpy_array_equal(
+            ci.isin(["c", "a", "b", np.nan]), np.array([True] * 6)
+        )
+
+        # mismatched categorical -> coerced to ndarray so doesn't matter
+        result = ci.isin(ci.set_categories(list("abcdefghi")))
+        expected = np.array([True] * 6)
+        tm.assert_numpy_array_equal(result, expected)
+
+        result = ci.isin(ci.set_categories(list("defghi")))
+        expected = np.array([False] * 5 + [True])
+        tm.assert_numpy_array_equal(result, expected)
+
+    def test_isin_overlapping_intervals(self):
+        # GH 34974
+        idx = pd.IntervalIndex([pd.Interval(0, 2), pd.Interval(0, 1)])
+        result = CategoricalIndex(idx).isin(idx)
+        expected = np.array([True, True])
+        tm.assert_numpy_array_equal(result, expected)
+
+    def test_identical(self):
+        ci1 = CategoricalIndex(["a", "b"], categories=["a", "b"], ordered=True)
+        ci2 = CategoricalIndex(["a", "b"], categories=["a", "b", "c"], ordered=True)
+        assert ci1.identical(ci1)
+        assert ci1.identical(ci1.copy())
+        assert not ci1.identical(ci2)
+
+    def test_ensure_copied_data(self):
+        # gh-12309: Check the "copy" argument of each
+        # Index.__new__ is honored.
+        #
+        # Must be tested separately from other indexes because
+        # self.values is not an ndarray.
+        index = CategoricalIndex(list("ab") * 5)
+
+        result = CategoricalIndex(index.values, copy=True)
+        tm.assert_index_equal(index, result)
+        assert not np.shares_memory(result._data._codes, index._data._codes)
+
+        result = CategoricalIndex(index.values, copy=False)
+        assert result._data._codes is index._data._codes
+
+
+class TestCategoricalIndex2:
+    def test_view_i8(self):
+        # GH#25464
+        ci = CategoricalIndex(list("ab") * 50)
+        msg = "When changing to a larger dtype, its size must be a divisor"
+        with pytest.raises(ValueError, match=msg):
+            ci.view("i8")
+        with pytest.raises(ValueError, match=msg):
+            ci._data.view("i8")
+
+        ci = ci[:-4]  # length divisible by 8
+
+        res = ci.view("i8")
+        expected = ci._data.codes.view("i8")
+        tm.assert_numpy_array_equal(res, expected)
+
+        cat = ci._data
+        tm.assert_numpy_array_equal(cat.view("i8"), expected)
+
+    @pytest.mark.parametrize(
+        "dtype, engine_type",
+        [
+            (np.int8, libindex.Int8Engine),
+            (np.int16, libindex.Int16Engine),
+            (np.int32, libindex.Int32Engine),
+            (np.int64, libindex.Int64Engine),
+        ],
+    )
+    def test_engine_type(self, dtype, engine_type):
+        if dtype != np.int64:
+            # num. of uniques required to push CategoricalIndex.codes to a
+            # dtype (128 categories required for .codes dtype to be int16 etc.)
+            num_uniques = {np.int8: 1, np.int16: 128, np.int32: 32768}[dtype]
+            ci = CategoricalIndex(range(num_uniques))
+        else:
+            # having 2**32 - 2**31 categories would be very memory-intensive,
+            # so we cheat a bit with the dtype
+            ci = CategoricalIndex(range(32768))  # == 2**16 - 2**(16 - 1)
+            arr = ci.values._ndarray.astype("int64")
+            NDArrayBacked.__init__(ci._data, arr, ci.dtype)
+        assert np.issubdtype(ci.codes.dtype, dtype)
+        assert isinstance(ci._engine, engine_type)
+
+    @pytest.mark.parametrize(
+        "func,op_name",
+        [
+            (lambda idx: idx - idx, "__sub__"),
+            (lambda idx: idx + idx, "__add__"),
+            (lambda idx: idx - ["a", "b"], "__sub__"),
+            (lambda idx: idx + ["a", "b"], "__add__"),
+            (lambda idx: ["a", "b"] - idx, "__rsub__"),
+            (lambda idx: ["a", "b"] + idx, "__radd__"),
+        ],
+    )
+    def test_disallow_addsub_ops(self, func, op_name):
+        # GH 10039
+        # set ops (+/-) raise TypeError
+        idx = Index(Categorical(["a", "b"]))
+        cat_or_list = "'(Categorical|list)' and '(Categorical|list)'"
+        msg = "|".join(
+            [
+                f"cannot perform {op_name} with this index type: CategoricalIndex",
+                "can only concatenate list",
+                rf"unsupported operand type\(s\) for [\+-]: {cat_or_list}",
+            ]
+        )
+        with pytest.raises(TypeError, match=msg):
+            func(idx)
+
+    def test_method_delegation(self):
+        ci = CategoricalIndex(list("aabbca"), categories=list("cabdef"))
+        result = ci.set_categories(list("cab"))
+        tm.assert_index_equal(
+            result, CategoricalIndex(list("aabbca"), categories=list("cab"))
+        )
+
+        ci = CategoricalIndex(list("aabbca"), categories=list("cab"))
+        result = ci.rename_categories(list("efg"))
+        tm.assert_index_equal(
+            result, CategoricalIndex(list("ffggef"), categories=list("efg"))
+        )
+
+        # GH18862 (let rename_categories take callables)
+        result = ci.rename_categories(lambda x: x.upper())
+        tm.assert_index_equal(
+            result, CategoricalIndex(list("AABBCA"), categories=list("CAB"))
+        )
+
+        ci = CategoricalIndex(list("aabbca"), categories=list("cab"))
+        result = ci.add_categories(["d"])
+        tm.assert_index_equal(
+            result, CategoricalIndex(list("aabbca"), categories=list("cabd"))
+        )
+
+        ci = CategoricalIndex(list("aabbca"), categories=list("cab"))
+        result = ci.remove_categories(["c"])
+        tm.assert_index_equal(
+            result,
+            CategoricalIndex(list("aabb") + [np.nan] + ["a"], categories=list("ab")),
+        )
+
+        ci = CategoricalIndex(list("aabbca"), categories=list("cabdef"))
+        result = ci.as_unordered()
+        tm.assert_index_equal(result, ci)
+
+        ci = CategoricalIndex(list("aabbca"), categories=list("cabdef"))
+        result = ci.as_ordered()
+        tm.assert_index_equal(
+            result,
+            CategoricalIndex(list("aabbca"), categories=list("cabdef"), ordered=True),
+        )
+
+        # invalid
+        msg = "cannot use inplace with CategoricalIndex"
+        with pytest.raises(ValueError, match=msg):
+            ci.set_categories(list("cab"), inplace=True)
+
+    def test_remove_maintains_order(self):
+        ci = CategoricalIndex(list("abcdda"), categories=list("abcd"))
+        result = ci.reorder_categories(["d", "c", "b", "a"], ordered=True)
+        tm.assert_index_equal(
+            result,
+            CategoricalIndex(list("abcdda"), categories=list("dcba"), ordered=True),
+        )
+        result = result.remove_categories(["c"])
+        tm.assert_index_equal(
+            result,
+            CategoricalIndex(
+                ["a", "b", np.nan, "d", "d", "a"], categories=list("dba"), ordered=True
+            ),
+        )

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_constructors.py

@@ -0,0 +1,142 @@
+import numpy as np
+import pytest
+
+from pandas import (
+    Categorical,
+    CategoricalDtype,
+    CategoricalIndex,
+    Index,
+)
+import pandas._testing as tm
+
+
+class TestCategoricalIndexConstructors:
+    def test_construction_disallows_scalar(self):
+        msg = "must be called with a collection of some kind"
+        with pytest.raises(TypeError, match=msg):
+            CategoricalIndex(data=1, categories=list("abcd"), ordered=False)
+        with pytest.raises(TypeError, match=msg):
+            CategoricalIndex(categories=list("abcd"), ordered=False)
+
+    def test_construction(self):
+        ci = CategoricalIndex(list("aabbca"), categories=list("abcd"), ordered=False)
+        categories = ci.categories
+
+        result = Index(ci)
+        tm.assert_index_equal(result, ci, exact=True)
+        assert not result.ordered
+
+        result = Index(ci.values)
+        tm.assert_index_equal(result, ci, exact=True)
+        assert not result.ordered
+
+        # empty
+        result = CategoricalIndex([], categories=categories)
+        tm.assert_index_equal(result.categories, Index(categories))
+        tm.assert_numpy_array_equal(result.codes, np.array([], dtype="int8"))
+        assert not result.ordered
+
+        # passing categories
+        result = CategoricalIndex(list("aabbca"), categories=categories)
+        tm.assert_index_equal(result.categories, Index(categories))
+        tm.assert_numpy_array_equal(
+            result.codes, np.array([0, 0, 1, 1, 2, 0], dtype="int8")
+        )
+
+        c = Categorical(list("aabbca"))
+        result = CategoricalIndex(c)
+        tm.assert_index_equal(result.categories, Index(list("abc")))
+        tm.assert_numpy_array_equal(
+            result.codes, np.array([0, 0, 1, 1, 2, 0], dtype="int8")
+        )
+        assert not result.ordered
+
+        result = CategoricalIndex(c, categories=categories)
+        tm.assert_index_equal(result.categories, Index(categories))
+        tm.assert_numpy_array_equal(
+            result.codes, np.array([0, 0, 1, 1, 2, 0], dtype="int8")
+        )
+        assert not result.ordered
+
+        ci = CategoricalIndex(c, categories=list("abcd"))
+        result = CategoricalIndex(ci)
+        tm.assert_index_equal(result.categories, Index(categories))
+        tm.assert_numpy_array_equal(
+            result.codes, np.array([0, 0, 1, 1, 2, 0], dtype="int8")
+        )
+        assert not result.ordered
+
+        result = CategoricalIndex(ci, categories=list("ab"))
+        tm.assert_index_equal(result.categories, Index(list("ab")))
+        tm.assert_numpy_array_equal(
+            result.codes, np.array([0, 0, 1, 1, -1, 0], dtype="int8")
+        )
+        assert not result.ordered
+
+        result = CategoricalIndex(ci, categories=list("ab"), ordered=True)
+        tm.assert_index_equal(result.categories, Index(list("ab")))
+        tm.assert_numpy_array_equal(
+            result.codes, np.array([0, 0, 1, 1, -1, 0], dtype="int8")
+        )
+        assert result.ordered
+
+        result = CategoricalIndex(ci, categories=list("ab"), ordered=True)
+        expected = CategoricalIndex(
+            ci, categories=list("ab"), ordered=True, dtype="category"
+        )
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # turn me to an Index
+        result = Index(np.array(ci))
+        assert isinstance(result, Index)
+        assert not isinstance(result, CategoricalIndex)
+
+    def test_construction_with_dtype(self):
+        # specify dtype
+        ci = CategoricalIndex(list("aabbca"), categories=list("abc"), ordered=False)
+
+        result = Index(np.array(ci), dtype="category")
+        tm.assert_index_equal(result, ci, exact=True)
+
+        result = Index(np.array(ci).tolist(), dtype="category")
+        tm.assert_index_equal(result, ci, exact=True)
+
+        # these are generally only equal when the categories are reordered
+        ci = CategoricalIndex(list("aabbca"), categories=list("cab"), ordered=False)
+
+        result = Index(np.array(ci), dtype="category").reorder_categories(ci.categories)
+        tm.assert_index_equal(result, ci, exact=True)
+
+        # make sure indexes are handled
+        idx = Index(range(3))
+        expected = CategoricalIndex([0, 1, 2], categories=idx, ordered=True)
+        result = CategoricalIndex(idx, categories=idx, ordered=True)
+        tm.assert_index_equal(result, expected, exact=True)
+
+    def test_construction_empty_with_bool_categories(self):
+        # see GH#22702
+        cat = CategoricalIndex([], categories=[True, False])
+        categories = sorted(cat.categories.tolist())
+        assert categories == [False, True]
+
+    def test_construction_with_categorical_dtype(self):
+        # construction with CategoricalDtype
+        # GH#18109
+        data, cats, ordered = "a a b b".split(), "c b a".split(), True
+        dtype = CategoricalDtype(categories=cats, ordered=ordered)
+
+        result = CategoricalIndex(data, dtype=dtype)
+        expected = CategoricalIndex(data, categories=cats, ordered=ordered)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # GH#19032
+        result = Index(data, dtype=dtype)
+        tm.assert_index_equal(result, expected, exact=True)
+
+        # error when combining categories/ordered and dtype kwargs
+        msg = "Cannot specify `categories` or `ordered` together with `dtype`."
+        with pytest.raises(ValueError, match=msg):
+            CategoricalIndex(data, categories=cats, dtype=dtype)
+
+        with pytest.raises(ValueError, match=msg):
+            CategoricalIndex(data, ordered=ordered, dtype=dtype)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_equals.py

@@ -0,0 +1,96 @@
+import numpy as np
+import pytest
+
+from pandas import (
+    Categorical,
+    CategoricalIndex,
+    Index,
+    MultiIndex,
+)
+
+
+class TestEquals:
+    def test_equals_categorical(self):
+        ci1 = CategoricalIndex(["a", "b"], categories=["a", "b"], ordered=True)
+        ci2 = CategoricalIndex(["a", "b"], categories=["a", "b", "c"], ordered=True)
+
+        assert ci1.equals(ci1)
+        assert not ci1.equals(ci2)
+        assert ci1.equals(ci1.astype(object))
+        assert ci1.astype(object).equals(ci1)
+
+        assert (ci1 == ci1).all()
+        assert not (ci1 != ci1).all()
+        assert not (ci1 > ci1).all()
+        assert not (ci1 < ci1).all()
+        assert (ci1 <= ci1).all()
+        assert (ci1 >= ci1).all()
+
+        assert not (ci1 == 1).all()
+        assert (ci1 == Index(["a", "b"])).all()
+        assert (ci1 == ci1.values).all()
+
+        # invalid comparisons
+        with pytest.raises(ValueError, match="Lengths must match"):
+            ci1 == Index(["a", "b", "c"])
+
+        msg = "Categoricals can only be compared if 'categories' are the same"
+        with pytest.raises(TypeError, match=msg):
+            ci1 == ci2
+        with pytest.raises(TypeError, match=msg):
+            ci1 == Categorical(ci1.values, ordered=False)
+        with pytest.raises(TypeError, match=msg):
+            ci1 == Categorical(ci1.values, categories=list("abc"))
+
+        # tests
+        # make sure that we are testing for category inclusion properly
+        ci = CategoricalIndex(list("aabca"), categories=["c", "a", "b"])
+        assert not ci.equals(list("aabca"))
+        # Same categories, but different order
+        # Unordered
+        assert ci.equals(CategoricalIndex(list("aabca")))
+        # Ordered
+        assert not ci.equals(CategoricalIndex(list("aabca"), ordered=True))
+        assert ci.equals(ci.copy())
+
+        ci = CategoricalIndex(list("aabca") + [np.nan], categories=["c", "a", "b"])
+        assert not ci.equals(list("aabca"))
+        assert not ci.equals(CategoricalIndex(list("aabca")))
+        assert ci.equals(ci.copy())
+
+        ci = CategoricalIndex(list("aabca") + [np.nan], categories=["c", "a", "b"])
+        assert not ci.equals(list("aabca") + [np.nan])
+        assert ci.equals(CategoricalIndex(list("aabca") + [np.nan]))
+        assert not ci.equals(CategoricalIndex(list("aabca") + [np.nan], ordered=True))
+        assert ci.equals(ci.copy())
+
+    def test_equals_categorical_unordered(self):
+        # https://github.com/pandas-dev/pandas/issues/16603
+        a = CategoricalIndex(["A"], categories=["A", "B"])
+        b = CategoricalIndex(["A"], categories=["B", "A"])
+        c = CategoricalIndex(["C"], categories=["B", "A"])
+        assert a.equals(b)
+        assert not a.equals(c)
+        assert not b.equals(c)
+
+    def test_equals_non_category(self):
+        # GH#37667 Case where other contains a value not among ci's
+        #  categories ("D") and also contains np.nan
+        ci = CategoricalIndex(["A", "B", np.nan, np.nan])
+        other = Index(["A", "B", "D", np.nan])
+
+        assert not ci.equals(other)
+
+    def test_equals_multiindex(self):
+        # dont raise NotImplementedError when calling is_dtype_compat
+
+        mi = MultiIndex.from_arrays([["A", "B", "C", "D"], range(4)])
+        ci = mi.to_flat_index().astype("category")
+
+        assert not ci.equals(mi)
+
+    def test_equals_string_dtype(self, any_string_dtype):
+        # GH#55364
+        idx = CategoricalIndex(list("abc"), name="B")
+        other = Index(["a", "b", "c"], name="B", dtype=any_string_dtype)
+        assert idx.equals(other)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_fillna.py

@@ -0,0 +1,54 @@
+import numpy as np
+import pytest
+
+from pandas import CategoricalIndex
+import pandas._testing as tm
+
+
+class TestFillNA:
+    def test_fillna_categorical(self):
+        # GH#11343
+        idx = CategoricalIndex([1.0, np.nan, 3.0, 1.0], name="x")
+        # fill by value in categories
+        exp = CategoricalIndex([1.0, 1.0, 3.0, 1.0], name="x")
+        tm.assert_index_equal(idx.fillna(1.0), exp)
+
+        cat = idx._data
+
+        # fill by value not in categories raises TypeError on EA, casts on CI
+        msg = "Cannot setitem on a Categorical with a new category"
+        with pytest.raises(TypeError, match=msg):
+            cat.fillna(2.0)
+
+        result = idx.fillna(2.0)
+        expected = idx.astype(object).fillna(2.0)
+        tm.assert_index_equal(result, expected)
+
+    def test_fillna_copies_with_no_nas(self):
+        # Nothing to fill, should still get a copy for the Categorical method,
+        #  but OK to get a view on CategoricalIndex method
+        ci = CategoricalIndex([0, 1, 1])
+        result = ci.fillna(0)
+        assert result is not ci
+        assert tm.shares_memory(result, ci)
+
+        # But at the EA level we always get a copy.
+        cat = ci._data
+        result = cat.fillna(0)
+        assert result._ndarray is not cat._ndarray
+        assert result._ndarray.base is None
+        assert not tm.shares_memory(result, cat)
+
+    def test_fillna_validates_with_no_nas(self):
+        # We validate the fill value even if fillna is a no-op
+        ci = CategoricalIndex([2, 3, 3])
+        cat = ci._data
+
+        msg = "Cannot setitem on a Categorical with a new category"
+        res = ci.fillna(False)
+        # nothing to fill, so we dont cast
+        tm.assert_index_equal(res, ci)
+
+        # Same check directly on the Categorical
+        with pytest.raises(TypeError, match=msg):
+            cat.fillna(False)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_formats.py

@@ -0,0 +1,120 @@
+"""
+Tests for CategoricalIndex.__repr__ and related methods.
+"""
+import pytest
+
+from pandas._config import using_pyarrow_string_dtype
+import pandas._config.config as cf
+
+from pandas import CategoricalIndex
+import pandas._testing as tm
+
+
+class TestCategoricalIndexRepr:
+    def test_format_different_scalar_lengths(self):
+        # GH#35439
+        idx = CategoricalIndex(["aaaaaaaaa", "b"])
+        expected = ["aaaaaaaaa", "b"]
+        msg = r"CategoricalIndex\.format is deprecated"
+        with tm.assert_produces_warning(FutureWarning, match=msg):
+            assert idx.format() == expected
+
+    @pytest.mark.xfail(using_pyarrow_string_dtype(), reason="repr different")
+    def test_string_categorical_index_repr(self):
+        # short
+        idx = CategoricalIndex(["a", "bb", "ccc"])
+        expected = """CategoricalIndex(['a', 'bb', 'ccc'], categories=['a', 'bb', 'ccc'], ordered=False, dtype='category')"""  # noqa: E501
+        assert repr(idx) == expected
+
+        # multiple lines
+        idx = CategoricalIndex(["a", "bb", "ccc"] * 10)
+        expected = """CategoricalIndex(['a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a',
+                  'bb', 'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a', 'bb',
+                  'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc'],
+                 categories=['a', 'bb', 'ccc'], ordered=False, dtype='category')"""  # noqa: E501
+
+        assert repr(idx) == expected
+
+        # truncated
+        idx = CategoricalIndex(["a", "bb", "ccc"] * 100)
+        expected = """CategoricalIndex(['a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a',
+                  ...
+                  'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc', 'a', 'bb', 'ccc'],
+                 categories=['a', 'bb', 'ccc'], ordered=False, dtype='category', length=300)"""  # noqa: E501
+
+        assert repr(idx) == expected
+
+        # larger categories
+        idx = CategoricalIndex(list("abcdefghijklmmo"))
+        expected = """CategoricalIndex(['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l',
+                  'm', 'm', 'o'],
+                 categories=['a', 'b', 'c', 'd', ..., 'k', 'l', 'm', 'o'], ordered=False, dtype='category')"""  # noqa: E501
+
+        assert repr(idx) == expected
+
+        # short
+        idx = CategoricalIndex(["あ", "いい", "ううう"])
+        expected = """CategoricalIndex(['あ', 'いい', 'ううう'], categories=['あ', 'いい', 'ううう'], ordered=False, dtype='category')"""  # noqa: E501
+        assert repr(idx) == expected
+
+        # multiple lines
+        idx = CategoricalIndex(["あ", "いい", "ううう"] * 10)
+        expected = """CategoricalIndex(['あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ',
+                  'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい',
+                  'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう'],
+                 categories=['あ', 'いい', 'ううう'], ordered=False, dtype='category')"""  # noqa: E501
+
+        assert repr(idx) == expected
+
+        # truncated
+        idx = CategoricalIndex(["あ", "いい", "ううう"] * 100)
+        expected = """CategoricalIndex(['あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ',
+                  ...
+                  'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう'],
+                 categories=['あ', 'いい', 'ううう'], ordered=False, dtype='category', length=300)"""  # noqa: E501
+
+        assert repr(idx) == expected
+
+        # larger categories
+        idx = CategoricalIndex(list("あいうえおかきくけこさしすせそ"))
+        expected = """CategoricalIndex(['あ', 'い', 'う', 'え', 'お', 'か', 'き', 'く', 'け', 'こ', 'さ', 'し',
+                  'す', 'せ', 'そ'],
+                 categories=['あ', 'い', 'う', 'え', ..., 'し', 'す', 'せ', 'そ'], ordered=False, dtype='category')"""  # noqa: E501
+
+        assert repr(idx) == expected
+
+        # Enable Unicode option -----------------------------------------
+        with cf.option_context("display.unicode.east_asian_width", True):
+            # short
+            idx = CategoricalIndex(["あ", "いい", "ううう"])
+            expected = """CategoricalIndex(['あ', 'いい', 'ううう'], categories=['あ', 'いい', 'ううう'], ordered=False, dtype='category')"""  # noqa: E501
+            assert repr(idx) == expected
+
+            # multiple lines
+            idx = CategoricalIndex(["あ", "いい", "ううう"] * 10)
+            expected = """CategoricalIndex(['あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい',
+                  'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう',
+                  'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい',
+                  'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう'],
+                 categories=['あ', 'いい', 'ううう'], ordered=False, dtype='category')"""  # noqa: E501
+
+            assert repr(idx) == expected
+
+            # truncated
+            idx = CategoricalIndex(["あ", "いい", "ううう"] * 100)
+            expected = """CategoricalIndex(['あ', 'いい', 'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい',
+                  'ううう', 'あ',
+                  ...
+                  'ううう', 'あ', 'いい', 'ううう', 'あ', 'いい', 'ううう',
+                  'あ', 'いい', 'ううう'],
+                 categories=['あ', 'いい', 'ううう'], ordered=False, dtype='category', length=300)"""  # noqa: E501
+
+            assert repr(idx) == expected
+
+            # larger categories
+            idx = CategoricalIndex(list("あいうえおかきくけこさしすせそ"))
+            expected = """CategoricalIndex(['あ', 'い', 'う', 'え', 'お', 'か', 'き', 'く', 'け', 'こ',
+                  'さ', 'し', 'す', 'せ', 'そ'],
+                 categories=['あ', 'い', 'う', 'え', ..., 'し', 'す', 'せ', 'そ'], ordered=False, dtype='category')"""  # noqa: E501
+
+            assert repr(idx) == expected

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_indexing.py

@@ -0,0 +1,420 @@
+import numpy as np
+import pytest
+
+from pandas.errors import InvalidIndexError
+
+import pandas as pd
+from pandas import (
+    CategoricalIndex,
+    Index,
+    IntervalIndex,
+    Timestamp,
+)
+import pandas._testing as tm
+
+
+class TestTake:
+    def test_take_fill_value(self):
+        # GH 12631
+
+        # numeric category
+        idx = CategoricalIndex([1, 2, 3], name="xxx")
+        result = idx.take(np.array([1, 0, -1]))
+        expected = CategoricalIndex([2, 1, 3], name="xxx")
+        tm.assert_index_equal(result, expected)
+        tm.assert_categorical_equal(result.values, expected.values)
+
+        # fill_value
+        result = idx.take(np.array([1, 0, -1]), fill_value=True)
+        expected = CategoricalIndex([2, 1, np.nan], categories=[1, 2, 3], name="xxx")
+        tm.assert_index_equal(result, expected)
+        tm.assert_categorical_equal(result.values, expected.values)
+
+        # allow_fill=False
+        result = idx.take(np.array([1, 0, -1]), allow_fill=False, fill_value=True)
+        expected = CategoricalIndex([2, 1, 3], name="xxx")
+        tm.assert_index_equal(result, expected)
+        tm.assert_categorical_equal(result.values, expected.values)
+
+        # object category
+        idx = CategoricalIndex(
+            list("CBA"), categories=list("ABC"), ordered=True, name="xxx"
+        )
+        result = idx.take(np.array([1, 0, -1]))
+        expected = CategoricalIndex(
+            list("BCA"), categories=list("ABC"), ordered=True, name="xxx"
+        )
+        tm.assert_index_equal(result, expected)
+        tm.assert_categorical_equal(result.values, expected.values)
+
+        # fill_value
+        result = idx.take(np.array([1, 0, -1]), fill_value=True)
+        expected = CategoricalIndex(
+            ["B", "C", np.nan], categories=list("ABC"), ordered=True, name="xxx"
+        )
+        tm.assert_index_equal(result, expected)
+        tm.assert_categorical_equal(result.values, expected.values)
+
+        # allow_fill=False
+        result = idx.take(np.array([1, 0, -1]), allow_fill=False, fill_value=True)
+        expected = CategoricalIndex(
+            list("BCA"), categories=list("ABC"), ordered=True, name="xxx"
+        )
+        tm.assert_index_equal(result, expected)
+        tm.assert_categorical_equal(result.values, expected.values)
+
+        msg = (
+            "When allow_fill=True and fill_value is not None, "
+            "all indices must be >= -1"
+        )
+        with pytest.raises(ValueError, match=msg):
+            idx.take(np.array([1, 0, -2]), fill_value=True)
+        with pytest.raises(ValueError, match=msg):
+            idx.take(np.array([1, 0, -5]), fill_value=True)
+
+        msg = "index -5 is out of bounds for (axis 0 with )?size 3"
+        with pytest.raises(IndexError, match=msg):
+            idx.take(np.array([1, -5]))
+
+    def test_take_fill_value_datetime(self):
+        # datetime category
+        idx = pd.DatetimeIndex(["2011-01-01", "2011-02-01", "2011-03-01"], name="xxx")
+        idx = CategoricalIndex(idx)
+        result = idx.take(np.array([1, 0, -1]))
+        expected = pd.DatetimeIndex(
+            ["2011-02-01", "2011-01-01", "2011-03-01"], name="xxx"
+        )
+        expected = CategoricalIndex(expected)
+        tm.assert_index_equal(result, expected)
+
+        # fill_value
+        result = idx.take(np.array([1, 0, -1]), fill_value=True)
+        expected = pd.DatetimeIndex(["2011-02-01", "2011-01-01", "NaT"], name="xxx")
+        exp_cats = pd.DatetimeIndex(["2011-01-01", "2011-02-01", "2011-03-01"])
+        expected = CategoricalIndex(expected, categories=exp_cats)
+        tm.assert_index_equal(result, expected)
+
+        # allow_fill=False
+        result = idx.take(np.array([1, 0, -1]), allow_fill=False, fill_value=True)
+        expected = pd.DatetimeIndex(
+            ["2011-02-01", "2011-01-01", "2011-03-01"], name="xxx"
+        )
+        expected = CategoricalIndex(expected)
+        tm.assert_index_equal(result, expected)
+
+        msg = (
+            "When allow_fill=True and fill_value is not None, "
+            "all indices must be >= -1"
+        )
+        with pytest.raises(ValueError, match=msg):
+            idx.take(np.array([1, 0, -2]), fill_value=True)
+        with pytest.raises(ValueError, match=msg):
+            idx.take(np.array([1, 0, -5]), fill_value=True)
+
+        msg = "index -5 is out of bounds for (axis 0 with )?size 3"
+        with pytest.raises(IndexError, match=msg):
+            idx.take(np.array([1, -5]))
+
+    def test_take_invalid_kwargs(self):
+        idx = CategoricalIndex([1, 2, 3], name="foo")
+        indices = [1, 0, -1]
+
+        msg = r"take\(\) got an unexpected keyword argument 'foo'"
+        with pytest.raises(TypeError, match=msg):
+            idx.take(indices, foo=2)
+
+        msg = "the 'out' parameter is not supported"
+        with pytest.raises(ValueError, match=msg):
+            idx.take(indices, out=indices)
+
+        msg = "the 'mode' parameter is not supported"
+        with pytest.raises(ValueError, match=msg):
+            idx.take(indices, mode="clip")
+
+
+class TestGetLoc:
+    def test_get_loc(self):
+        # GH 12531
+        cidx1 = CategoricalIndex(list("abcde"), categories=list("edabc"))
+        idx1 = Index(list("abcde"))
+        assert cidx1.get_loc("a") == idx1.get_loc("a")
+        assert cidx1.get_loc("e") == idx1.get_loc("e")
+
+        for i in [cidx1, idx1]:
+            with pytest.raises(KeyError, match="'NOT-EXIST'"):
+                i.get_loc("NOT-EXIST")
+
+        # non-unique
+        cidx2 = CategoricalIndex(list("aacded"), categories=list("edabc"))
+        idx2 = Index(list("aacded"))
+
+        # results in bool array
+        res = cidx2.get_loc("d")
+        tm.assert_numpy_array_equal(res, idx2.get_loc("d"))
+        tm.assert_numpy_array_equal(
+            res, np.array([False, False, False, True, False, True])
+        )
+        # unique element results in scalar
+        res = cidx2.get_loc("e")
+        assert res == idx2.get_loc("e")
+        assert res == 4
+
+        for i in [cidx2, idx2]:
+            with pytest.raises(KeyError, match="'NOT-EXIST'"):
+                i.get_loc("NOT-EXIST")
+
+        # non-unique, sliceable
+        cidx3 = CategoricalIndex(list("aabbb"), categories=list("abc"))
+        idx3 = Index(list("aabbb"))
+
+        # results in slice
+        res = cidx3.get_loc("a")
+        assert res == idx3.get_loc("a")
+        assert res == slice(0, 2, None)
+
+        res = cidx3.get_loc("b")
+        assert res == idx3.get_loc("b")
+        assert res == slice(2, 5, None)
+
+        for i in [cidx3, idx3]:
+            with pytest.raises(KeyError, match="'c'"):
+                i.get_loc("c")
+
+    def test_get_loc_unique(self):
+        cidx = CategoricalIndex(list("abc"))
+        result = cidx.get_loc("b")
+        assert result == 1
+
+    def test_get_loc_monotonic_nonunique(self):
+        cidx = CategoricalIndex(list("abbc"))
+        result = cidx.get_loc("b")
+        expected = slice(1, 3, None)
+        assert result == expected
+
+    def test_get_loc_nonmonotonic_nonunique(self):
+        cidx = CategoricalIndex(list("abcb"))
+        result = cidx.get_loc("b")
+        expected = np.array([False, True, False, True], dtype=bool)
+        tm.assert_numpy_array_equal(result, expected)
+
+    def test_get_loc_nan(self):
+        # GH#41933
+        ci = CategoricalIndex(["A", "B", np.nan])
+        res = ci.get_loc(np.nan)
+
+        assert res == 2
+
+
+class TestGetIndexer:
+    def test_get_indexer_base(self):
+        # Determined by cat ordering.
+        idx = CategoricalIndex(list("cab"), categories=list("cab"))
+        expected = np.arange(len(idx), dtype=np.intp)
+
+        actual = idx.get_indexer(idx)
+        tm.assert_numpy_array_equal(expected, actual)
+
+        with pytest.raises(ValueError, match="Invalid fill method"):
+            idx.get_indexer(idx, method="invalid")
+
+    def test_get_indexer_requires_unique(self):
+        ci = CategoricalIndex(list("aabbca"), categories=list("cab"), ordered=False)
+        oidx = Index(np.array(ci))
+
+        msg = "Reindexing only valid with uniquely valued Index objects"
+
+        for n in [1, 2, 5, len(ci)]:
+            finder = oidx[np.random.default_rng(2).integers(0, len(ci), size=n)]
+
+            with pytest.raises(InvalidIndexError, match=msg):
+                ci.get_indexer(finder)
+
+        # see gh-17323
+        #
+        # Even when indexer is equal to the
+        # members in the index, we should
+        # respect duplicates instead of taking
+        # the fast-track path.
+        for finder in [list("aabbca"), list("aababca")]:
+            with pytest.raises(InvalidIndexError, match=msg):
+                ci.get_indexer(finder)
+
+    def test_get_indexer_non_unique(self):
+        idx1 = CategoricalIndex(list("aabcde"), categories=list("edabc"))
+        idx2 = CategoricalIndex(list("abf"))
+
+        for indexer in [idx2, list("abf"), Index(list("abf"))]:
+            msg = "Reindexing only valid with uniquely valued Index objects"
+            with pytest.raises(InvalidIndexError, match=msg):
+                idx1.get_indexer(indexer)
+
+            r1, _ = idx1.get_indexer_non_unique(indexer)
+            expected = np.array([0, 1, 2, -1], dtype=np.intp)
+            tm.assert_almost_equal(r1, expected)
+
+    def test_get_indexer_method(self):
+        idx1 = CategoricalIndex(list("aabcde"), categories=list("edabc"))
+        idx2 = CategoricalIndex(list("abf"))
+
+        msg = "method pad not yet implemented for CategoricalIndex"
+        with pytest.raises(NotImplementedError, match=msg):
+            idx2.get_indexer(idx1, method="pad")
+        msg = "method backfill not yet implemented for CategoricalIndex"
+        with pytest.raises(NotImplementedError, match=msg):
+            idx2.get_indexer(idx1, method="backfill")
+
+        msg = "method nearest not yet implemented for CategoricalIndex"
+        with pytest.raises(NotImplementedError, match=msg):
+            idx2.get_indexer(idx1, method="nearest")
+
+    def test_get_indexer_array(self):
+        arr = np.array(
+            [Timestamp("1999-12-31 00:00:00"), Timestamp("2000-12-31 00:00:00")],
+            dtype=object,
+        )
+        cats = [Timestamp("1999-12-31 00:00:00"), Timestamp("2000-12-31 00:00:00")]
+        ci = CategoricalIndex(cats, categories=cats, ordered=False, dtype="category")
+        result = ci.get_indexer(arr)
+        expected = np.array([0, 1], dtype="intp")
+        tm.assert_numpy_array_equal(result, expected)
+
+    def test_get_indexer_same_categories_same_order(self):
+        ci = CategoricalIndex(["a", "b"], categories=["a", "b"])
+
+        result = ci.get_indexer(CategoricalIndex(["b", "b"], categories=["a", "b"]))
+        expected = np.array([1, 1], dtype="intp")
+        tm.assert_numpy_array_equal(result, expected)
+
+    def test_get_indexer_same_categories_different_order(self):
+        # https://github.com/pandas-dev/pandas/issues/19551
+        ci = CategoricalIndex(["a", "b"], categories=["a", "b"])
+
+        result = ci.get_indexer(CategoricalIndex(["b", "b"], categories=["b", "a"]))
+        expected = np.array([1, 1], dtype="intp")
+        tm.assert_numpy_array_equal(result, expected)
+
+    def test_get_indexer_nans_in_index_and_target(self):
+        # GH 45361
+        ci = CategoricalIndex([1, 2, np.nan, 3])
+        other1 = [2, 3, 4, np.nan]
+        res1 = ci.get_indexer(other1)
+        expected1 = np.array([1, 3, -1, 2], dtype=np.intp)
+        tm.assert_numpy_array_equal(res1, expected1)
+        other2 = [1, 4, 2, 3]
+        res2 = ci.get_indexer(other2)
+        expected2 = np.array([0, -1, 1, 3], dtype=np.intp)
+        tm.assert_numpy_array_equal(res2, expected2)
+
+
+class TestWhere:
+    def test_where(self, listlike_box):
+        klass = listlike_box
+
+        i = CategoricalIndex(list("aabbca"), categories=list("cab"), ordered=False)
+        cond = [True] * len(i)
+        expected = i
+        result = i.where(klass(cond))
+        tm.assert_index_equal(result, expected)
+
+        cond = [False] + [True] * (len(i) - 1)
+        expected = CategoricalIndex([np.nan] + i[1:].tolist(), categories=i.categories)
+        result = i.where(klass(cond))
+        tm.assert_index_equal(result, expected)
+
+    def test_where_non_categories(self):
+        ci = CategoricalIndex(["a", "b", "c", "d"])
+        mask = np.array([True, False, True, False])
+
+        result = ci.where(mask, 2)
+        expected = Index(["a", 2, "c", 2], dtype=object)
+        tm.assert_index_equal(result, expected)
+
+        msg = "Cannot setitem on a Categorical with a new category"
+        with pytest.raises(TypeError, match=msg):
+            # Test the Categorical method directly
+            ci._data._where(mask, 2)
+
+
+class TestContains:
+    def test_contains(self):
+        ci = CategoricalIndex(list("aabbca"), categories=list("cabdef"), ordered=False)
+
+        assert "a" in ci
+        assert "z" not in ci
+        assert "e" not in ci
+        assert np.nan not in ci
+
+        # assert codes NOT in index
+        assert 0 not in ci
+        assert 1 not in ci
+
+    def test_contains_nan(self):
+        ci = CategoricalIndex(list("aabbca") + [np.nan], categories=list("cabdef"))
+        assert np.nan in ci
+
+    @pytest.mark.parametrize("unwrap", [True, False])
+    def test_contains_na_dtype(self, unwrap):
+        dti = pd.date_range("2016-01-01", periods=100).insert(0, pd.NaT)
+        pi = dti.to_period("D")
+        tdi = dti - dti[-1]
+        ci = CategoricalIndex(dti)
+
+        obj = ci
+        if unwrap:
+            obj = ci._data
+
+        assert np.nan in obj
+        assert None in obj
+        assert pd.NaT in obj
+        assert np.datetime64("NaT") in obj
+        assert np.timedelta64("NaT") not in obj
+
+        obj2 = CategoricalIndex(tdi)
+        if unwrap:
+            obj2 = obj2._data
+
+        assert np.nan in obj2
+        assert None in obj2
+        assert pd.NaT in obj2
+        assert np.datetime64("NaT") not in obj2
+        assert np.timedelta64("NaT") in obj2
+
+        obj3 = CategoricalIndex(pi)
+        if unwrap:
+            obj3 = obj3._data
+
+        assert np.nan in obj3
+        assert None in obj3
+        assert pd.NaT in obj3
+        assert np.datetime64("NaT") not in obj3
+        assert np.timedelta64("NaT") not in obj3
+
+    @pytest.mark.parametrize(
+        "item, expected",
+        [
+            (pd.Interval(0, 1), True),
+            (1.5, True),
+            (pd.Interval(0.5, 1.5), False),
+            ("a", False),
+            (Timestamp(1), False),
+            (pd.Timedelta(1), False),
+        ],
+        ids=str,
+    )
+    def test_contains_interval(self, item, expected):
+        # GH 23705
+        ci = CategoricalIndex(IntervalIndex.from_breaks(range(3)))
+        result = item in ci
+        assert result is expected
+
+    def test_contains_list(self):
+        # GH#21729
+        idx = CategoricalIndex([1, 2, 3])
+
+        assert "a" not in idx
+
+        with pytest.raises(TypeError, match="unhashable type"):
+            ["a"] in idx
+
+        with pytest.raises(TypeError, match="unhashable type"):
+            ["a", "b"] in idx

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_map.py

@@ -0,0 +1,144 @@
+import numpy as np
+import pytest
+
+import pandas as pd
+from pandas import (
+    CategoricalIndex,
+    Index,
+    Series,
+)
+import pandas._testing as tm
+
+
+@pytest.mark.parametrize(
+    "data, categories",
+    [
+        (list("abcbca"), list("cab")),
+        (pd.interval_range(0, 3).repeat(3), pd.interval_range(0, 3)),
+    ],
+    ids=["string", "interval"],
+)
+def test_map_str(data, categories, ordered):
+    # GH 31202 - override base class since we want to maintain categorical/ordered
+    index = CategoricalIndex(data, categories=categories, ordered=ordered)
+    result = index.map(str)
+    expected = CategoricalIndex(
+        map(str, data), categories=map(str, categories), ordered=ordered
+    )
+    tm.assert_index_equal(result, expected)
+
+
+def test_map():
+    ci = CategoricalIndex(list("ABABC"), categories=list("CBA"), ordered=True)
+    result = ci.map(lambda x: x.lower())
+    exp = CategoricalIndex(list("ababc"), categories=list("cba"), ordered=True)
+    tm.assert_index_equal(result, exp)
+
+    ci = CategoricalIndex(
+        list("ABABC"), categories=list("BAC"), ordered=False, name="XXX"
+    )
+    result = ci.map(lambda x: x.lower())
+    exp = CategoricalIndex(
+        list("ababc"), categories=list("bac"), ordered=False, name="XXX"
+    )
+    tm.assert_index_equal(result, exp)
+
+    # GH 12766: Return an index not an array
+    tm.assert_index_equal(
+        ci.map(lambda x: 1), Index(np.array([1] * 5, dtype=np.int64), name="XXX")
+    )
+
+    # change categories dtype
+    ci = CategoricalIndex(list("ABABC"), categories=list("BAC"), ordered=False)
+
+    def f(x):
+        return {"A": 10, "B": 20, "C": 30}.get(x)
+
+    result = ci.map(f)
+    exp = CategoricalIndex([10, 20, 10, 20, 30], categories=[20, 10, 30], ordered=False)
+    tm.assert_index_equal(result, exp)
+
+    result = ci.map(Series([10, 20, 30], index=["A", "B", "C"]))
+    tm.assert_index_equal(result, exp)
+
+    result = ci.map({"A": 10, "B": 20, "C": 30})
+    tm.assert_index_equal(result, exp)
+
+
+def test_map_with_categorical_series():
+    # GH 12756
+    a = Index([1, 2, 3, 4])
+    b = Series(["even", "odd", "even", "odd"], dtype="category")
+    c = Series(["even", "odd", "even", "odd"])
+
+    exp = CategoricalIndex(["odd", "even", "odd", np.nan])
+    tm.assert_index_equal(a.map(b), exp)
+    exp = Index(["odd", "even", "odd", np.nan])
+    tm.assert_index_equal(a.map(c), exp)
+
+
+@pytest.mark.parametrize(
+    ("data", "f", "expected"),
+    (
+        ([1, 1, np.nan], pd.isna, CategoricalIndex([False, False, np.nan])),
+        ([1, 2, np.nan], pd.isna, Index([False, False, np.nan])),
+        ([1, 1, np.nan], {1: False}, CategoricalIndex([False, False, np.nan])),
+        ([1, 2, np.nan], {1: False, 2: False}, Index([False, False, np.nan])),
+        (
+            [1, 1, np.nan],
+            Series([False, False]),
+            CategoricalIndex([False, False, np.nan]),
+        ),
+        (
+            [1, 2, np.nan],
+            Series([False, False, False]),
+            Index([False, False, np.nan]),
+        ),
+    ),
+)
+def test_map_with_nan_ignore(data, f, expected):  # GH 24241
+    values = CategoricalIndex(data)
+    result = values.map(f, na_action="ignore")
+    tm.assert_index_equal(result, expected)
+
+
+@pytest.mark.parametrize(
+    ("data", "f", "expected"),
+    (
+        ([1, 1, np.nan], pd.isna, Index([False, False, True])),
+        ([1, 2, np.nan], pd.isna, Index([False, False, True])),
+        ([1, 1, np.nan], {1: False}, CategoricalIndex([False, False, np.nan])),
+        ([1, 2, np.nan], {1: False, 2: False}, Index([False, False, np.nan])),
+        (
+            [1, 1, np.nan],
+            Series([False, False]),
+            CategoricalIndex([False, False, np.nan]),
+        ),
+        (
+            [1, 2, np.nan],
+            Series([False, False, False]),
+            Index([False, False, np.nan]),
+        ),
+    ),
+)
+def test_map_with_nan_none(data, f, expected):  # GH 24241
+    values = CategoricalIndex(data)
+    result = values.map(f, na_action=None)
+    tm.assert_index_equal(result, expected)
+
+
+def test_map_with_dict_or_series():
+    orig_values = ["a", "B", 1, "a"]
+    new_values = ["one", 2, 3.0, "one"]
+    cur_index = CategoricalIndex(orig_values, name="XXX")
+    expected = CategoricalIndex(new_values, name="XXX", categories=[3.0, 2, "one"])
+
+    mapper = Series(new_values[:-1], index=orig_values[:-1])
+    result = cur_index.map(mapper)
+    # Order of categories in result can be different
+    tm.assert_index_equal(result, expected)
+
+    mapper = dict(zip(orig_values[:-1], new_values[:-1]))
+    result = cur_index.map(mapper)
+    # Order of categories in result can be different
+    tm.assert_index_equal(result, expected)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_reindex.py

@@ -0,0 +1,78 @@
+import numpy as np
+import pytest
+
+from pandas import (
+    Categorical,
+    CategoricalIndex,
+    Index,
+    Interval,
+)
+import pandas._testing as tm
+
+
+class TestReindex:
+    def test_reindex_list_non_unique(self):
+        # GH#11586
+        msg = "cannot reindex on an axis with duplicate labels"
+        ci = CategoricalIndex(["a", "b", "c", "a"])
+        with pytest.raises(ValueError, match=msg):
+            ci.reindex(["a", "c"])
+
+    def test_reindex_categorical_non_unique(self):
+        msg = "cannot reindex on an axis with duplicate labels"
+        ci = CategoricalIndex(["a", "b", "c", "a"])
+        with pytest.raises(ValueError, match=msg):
+            ci.reindex(Categorical(["a", "c"]))
+
+    def test_reindex_list_non_unique_unused_category(self):
+        msg = "cannot reindex on an axis with duplicate labels"
+        ci = CategoricalIndex(["a", "b", "c", "a"], categories=["a", "b", "c", "d"])
+        with pytest.raises(ValueError, match=msg):
+            ci.reindex(["a", "c"])
+
+    def test_reindex_categorical_non_unique_unused_category(self):
+        msg = "cannot reindex on an axis with duplicate labels"
+        ci = CategoricalIndex(["a", "b", "c", "a"], categories=["a", "b", "c", "d"])
+        with pytest.raises(ValueError, match=msg):
+            ci.reindex(Categorical(["a", "c"]))
+
+    def test_reindex_duplicate_target(self):
+        # See GH25459
+        cat = CategoricalIndex(["a", "b", "c"], categories=["a", "b", "c", "d"])
+        res, indexer = cat.reindex(["a", "c", "c"])
+        exp = Index(["a", "c", "c"])
+        tm.assert_index_equal(res, exp, exact=True)
+        tm.assert_numpy_array_equal(indexer, np.array([0, 2, 2], dtype=np.intp))
+
+        res, indexer = cat.reindex(
+            CategoricalIndex(["a", "c", "c"], categories=["a", "b", "c", "d"])
+        )
+        exp = CategoricalIndex(["a", "c", "c"], categories=["a", "b", "c", "d"])
+        tm.assert_index_equal(res, exp, exact=True)
+        tm.assert_numpy_array_equal(indexer, np.array([0, 2, 2], dtype=np.intp))
+
+    def test_reindex_empty_index(self):
+        # See GH16770
+        c = CategoricalIndex([])
+        res, indexer = c.reindex(["a", "b"])
+        tm.assert_index_equal(res, Index(["a", "b"]), exact=True)
+        tm.assert_numpy_array_equal(indexer, np.array([-1, -1], dtype=np.intp))
+
+    def test_reindex_categorical_added_category(self):
+        # GH 42424
+        ci = CategoricalIndex(
+            [Interval(0, 1, closed="right"), Interval(1, 2, closed="right")],
+            ordered=True,
+        )
+        ci_add = CategoricalIndex(
+            [
+                Interval(0, 1, closed="right"),
+                Interval(1, 2, closed="right"),
+                Interval(2, 3, closed="right"),
+                Interval(3, 4, closed="right"),
+            ],
+            ordered=True,
+        )
+        result, _ = ci.reindex(ci_add)
+        expected = ci_add
+        tm.assert_index_equal(expected, result)

lib/python3.12/site-packages/pandas/tests/indexes/categorical/test_setops.py

@@ -0,0 +1,18 @@
+import numpy as np
+import pytest
+
+from pandas import (
+    CategoricalIndex,
+    Index,
+)
+import pandas._testing as tm
+
+
+@pytest.mark.parametrize("na_value", [None, np.nan])
+def test_difference_with_na(na_value):
+    # GH 57318
+    ci = CategoricalIndex(["a", "b", "c", None])
+    other = Index(["c", na_value])
+    result = ci.difference(other)
+    expected = CategoricalIndex(["a", "b"], categories=["a", "b", "c"])
+    tm.assert_index_equal(result, expected)

lib/python3.12/site-packages/pandas/tests/indexes/conftest.py

@@ -0,0 +1,41 @@
+import numpy as np
+import pytest
+
+from pandas import (
+    Series,
+    array,
+)
+
+
+@pytest.fixture(params=[None, False])
+def sort(request):
+    """
+    Valid values for the 'sort' parameter used in the Index
+    setops methods (intersection, union, etc.)
+
+    Caution:
+        Don't confuse this one with the "sort" fixture used
+        for DataFrame.append or concat. That one has
+        parameters [True, False].
+
+        We can't combine them as sort=True is not permitted
+        in the Index setops methods.
+    """
+    return request.param
+
+
+@pytest.fixture(params=["D", "3D", "-3D", "h", "2h", "-2h", "min", "2min", "s", "-3s"])
+def freq_sample(request):
+    """
+    Valid values for 'freq' parameter used to create date_range and
+    timedelta_range..
+    """
+    return request.param
+
+
+@pytest.fixture(params=[list, tuple, np.array, array, Series])
+def listlike_box(request):
+    """
+    Types that may be passed as the indexer to searchsorted.
+    """
+    return request.param