Validation of optional regex columns in filtered dataframes can bypass checks about pandera HOT 5 CLOSED

df2 = pd.DataFrame(
    {
        "sample_id": ["a", "b", "c", "d", "e", "f", "g", "h", "i", "j"],
        "Cu_ppm": [10, 20, -1, 40, 50, 60, 70, 80, 90, 100],
        "Cs_ppb": [100, 200, 300, 400, 500, 600, 700, 800, 900, 1000],
        "spam": [100, 200, 300, 400, 500, 600, 700, 800, 900, 1000],
    }
)


schema7 = pa.DataFrameSchema(
    {
        r"^[A-Z][a-z]*_ppm$": pa.Column(
            float,
            nullable=True,
            coerce=True,
            checks=pa.Check.in_range(0, 1e6),
            required=False,
            regex=True,
        ),
    },
    index=pa.Index(str, nullable=False, unique=True, coerce=True),
    checks=valid_column_labels,
    strict='filter',
)


schema7.validate(df2)

valid_column_labels = [
    pa.Check(lambda df: df.columns.is_unique, element_wise=False, error="Columns must be unique"),
    pa.Check(
        lambda df: df.columns.str.split("_").str[0].isin(valid_prefixes).all(),
        element_wise=False,
        error="Element symbols must be valid",
    ),
]