KaneAI - World’s First E2E Software Testing Agent.Book a Demo Join TestMu Conference

How to use infer_series_statistics method in pandera

Best Python code snippet using pandera_python

test_schema_statistics.py

Source:test_schema_statistics.py

...202    ],203)204def test_infer_series_schema_statistics(series, expectation) -> None:205    """Test series statistics are correctly inferred."""206    statistics = schema_statistics.infer_series_statistics(series)207    _test_statistics(statistics, expectation)208@pytest.mark.parametrize(209    "null_index, series, expectation",210    [211        *[212            [213                0,214                pd.Series([1, 2, 3], dtype=str(data_type)),215                {216                    # introducing nans to integer arrays upcasts to float217                    "dtype": DEFAULT_FLOAT,218                    "nullable": True,219                    "checks": {220                        "greater_than_or_equal_to": 2,221                        "less_than_or_equal_to": 3,222                    },223                    "name": None,224                },225            ]226            for data_type in INTEGER_TYPES227        ],228        [229            # introducing nans to bool arrays upcasts to float except230            # for pandas >= 1.3.0231            0,232            pd.Series([True, False, True, False]),233            {234                "dtype": (235                    pandas_engine.Engine.dtype(pa.Object)236                    if pa.PANDAS_1_3_0_PLUS237                    else DEFAULT_FLOAT238                ),239                "nullable": True,240                "checks": (241                    None242                    if pa.PANDAS_1_3_0_PLUS243                    else {244                        "greater_than_or_equal_to": 0,245                        "less_than_or_equal_to": 1,246                    }247                ),248                "name": None,249            },250        ],251        [252            0,253            pd.Series(["a", "b", "c", "a"], dtype="category"),254            {255                "dtype": pandas_engine.Engine.dtype(pa.Category),256                "nullable": True,257                "checks": {"isin": ["a", "b", "c"]},258                "name": None,259            },260        ],261        [262            0,263            pd.Series(["a", "b", "c", "a"], name="str_series"),264            {265                "dtype": pandas_engine.Engine.dtype(str),266                "nullable": True,267                "checks": None,268                "name": "str_series",269            },270        ],271        [272            2,273            pd.Series(pd.to_datetime(["20180101", "20180102", "20180103"])),274            {275                "dtype": pandas_engine.Engine.dtype(pa.DateTime),276                "nullable": True,277                "checks": {278                    "greater_than_or_equal_to": pd.Timestamp("20180101"),279                    "less_than_or_equal_to": pd.Timestamp("20180102"),280                },281                "name": None,282            },283        ],284    ],285)286def test_infer_nullable_series_schema_statistics(287    null_index, series, expectation288):289    """Test nullable series statistics are correctly inferred."""290    series.iloc[null_index] = None291    statistics = schema_statistics.infer_series_statistics(series)292    _test_statistics(statistics, expectation)293@pytest.mark.parametrize(294    "null_values, dtype",295    (296        ([pd.NaT, pd.NaT], pandas_engine.Engine.dtype(pa.DateTime)),297        ([np.nan, np.nan], DEFAULT_FLOAT),298        ([None, None], pandas_engine.Engine.dtype(pa.Object)),299        # Mixed 'null' types300        ([pd.NaT, np.nan], pandas_engine.Engine.dtype(pa.DateTime)),301        ([None, pd.NaT], pandas_engine.Engine.dtype(pa.DateTime)),302        ([None, np.nan], DEFAULT_FLOAT),303    ),304)305def test_empty_series_schema_statistics(null_values, dtype):306    """Test 'empty' series statistics are correctly inferred."""307    series = pd.Series(null_values)308    statistics = schema_statistics.infer_series_statistics(series)309    expectation = {310        "dtype": dtype,311        "nullable": True,312        "checks": None,313        "name": None,314    }315    _test_statistics(statistics, expectation)316@pytest.mark.parametrize(317    "index, expectation",318    [319        [320            pd.RangeIndex(20),321            [322                {...

schema_inference.py

Source:schema_inference.py

...64    """Infer a SeriesSchema from a pandas DataFrame.65    :param series: Series object to infer.66    :returns: SeriesSchema67    """68    series_statistics = infer_series_statistics(series)69    schema = SeriesSchema(70        dtype=series_statistics["dtype"],71        checks=parse_check_statistics(series_statistics["checks"]),72        nullable=series_statistics["nullable"],73        name=series_statistics["name"],74        coerce=True,75    )76    schema._is_inferred = True...

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.