Function test_describe

dask/dataframe/tests/test_dataframe.py:401–460

(include, exclude, percentiles, subset)

    ],
)
def test_describe(include, exclude, percentiles, subset):
    data = {
        "a": ["aaa", "bbb", "bbb", None, None, "zzz"] * 2,
        "c": [None, 0, 1, 2, 3, 4] * 2,
        "d": [None, 0, 1] * 4,
        "e": [
            pd.Timestamp("2017-05-09 00:00:00.006000"),
            pd.Timestamp("2017-05-09 00:00:00.006000"),
            pd.Timestamp("2017-05-09 07:56:23.858694"),
            pd.Timestamp("2017-05-09 05:59:58.938999"),
            None,
            None,
        ]
        * 2,
        "f": [
            np.timedelta64(3, "D"),
            np.timedelta64(1, "D"),
            None,
            None,
            np.timedelta64(3, "D"),
            np.timedelta64(1, "D"),
        ]
        * 2,
        "g": [True, False, True] * 4,
    }

    # Arrange
    df = pd.DataFrame(data)
    df["a"] = df["a"].astype(get_string_dtype())

    if subset is not None:
        df = df.loc[:, subset]

    ddf = dd.from_pandas(df, 2)

    # Act
    actual = ddf.describe(
        include=include,
        exclude=exclude,
        percentiles=percentiles,
    )
    expected = df.describe(
        include=include,
        exclude=exclude,
        percentiles=percentiles,
    )

    if "e" in expected:
        expected = _drop_mean(expected, "e")

    assert_eq(actual, expected)

    # Check series
    if subset is None:
        for col in ["a", "c", "e", "g"]:
            expected = df[col].describe(include=include, exclude=exclude)
            if col == "e":
                expected = _drop_mean(expected)