Dataframe

Helpers for dealing with pandas.DataFrames

_fix_integer_dtypes(df)

Since some version, pandas can not infer in assert_frame_equals Int64 as int64

Parameters:

Name	Type	Description	Default
df	DataFrame	A dataframe, that will behave all int types as int64	required

Source code in bquest/dataframe.py

def _fix_integer_dtypes(df: pandas.DataFrame) -> None:
    """Since some version, pandas can not infer in assert_frame_equals Int64 as int64

    Args:
        df: A dataframe, that will behave all int types as int64
    """
    df[df.select_dtypes("Int64").columns] = df[df.select_dtypes("Int64").columns].astype("Int64")

assert_frame_equal(left, right, **kwargs)

Asserts that two dataframes are equal regardless of their order of rows

Parameters:

Name	Type	Description	Default
left	DataFrame	A dataframe, usually the result of a function under test	required
right	DataFrame	Another dataframe, usually what we expect in a test	required
**kwargs	Any	Keyword arguments of pandas.testing.assert_frame_equal https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.testing.assert_frame_equal.html	{}

Source code in bquest/dataframe.py

def assert_frame_equal(left: pandas.DataFrame, right: pandas.DataFrame, **kwargs: Any) -> None:
    """Asserts that two dataframes are equal regardless of their order of rows

    Args:
        left: A dataframe, usually the result of a function under test
        right: Another dataframe, usually what we expect in a test
        **kwargs: Keyword arguments of pandas.testing.assert_frame_equal <https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.testing.assert_frame_equal.html>
    """

    _fix_integer_dtypes(left)
    _fix_integer_dtypes(right)

    left_sorted = left[sorted(left.columns)].sort_values(sorted(left.columns)).reset_index(drop=True)
    right_sorted = right[sorted(right.columns)].sort_values(sorted(right.columns)).reset_index(drop=True)

    pd_test.assert_frame_equal(left_sorted, right_sorted, **kwargs)

standardize_frame_numerics(df, float_precision=2)

Standardizes numerics inside a dataframe to facilitate comparison between dataframes with respect to meaningful differences.

Parameters:

Name	Type	Description	Default
df	DataFrame	Pandas dataframe to be standardized	required
float_precision	int	level of precision for rounding floats	2

Returns:

Type	Description
DataFrame	Standardized dataframe

Source code in bquest/dataframe.py

def standardize_frame_numerics(df: pandas.DataFrame, float_precision: int = 2) -> pandas.DataFrame:
    """Standardizes numerics inside a dataframe to facilitate comparison between
     dataframes with respect to meaningful differences.

    Args:
        df: Pandas dataframe to be standardized
        float_precision: level of precision for rounding floats

    Returns:
        Standardized dataframe
    """
    df = df.round(float_precision)

    integer_columns = df.select_dtypes(POSSIBLE_INTEGER_DTYPES).columns

    for col in integer_columns:
        df[col] = df[col].astype(float)

    return df.fillna(value=np.nan).reset_index(drop=True)