pandas.to_numeric转化数据为数字型

to_numeric(arg, errors='raise', downcast=None)

    Convert argument to a numeric type.

    

    Parameters

    ----------

    arg : list, tuple, 1-d array, or Series

    errors : {'ignore', 'raise', 'coerce'}, default 'raise'

        - If 'raise', then invalid parsing will raise an exception

        - If 'coerce', then invalid parsing will be set as NaN

        - If 'ignore', then invalid parsing will return the input

    downcast : {'integer', 'signed', 'unsigned', 'float'} , default None

        If not None, and if the data has been successfully cast to a

        numerical dtype (or if the data was numeric to begin with),

        downcast that resulting data to the smallest numerical dtype

        possible according to the following rules:

    

        - 'integer' or 'signed': smallest signed int dtype (min.: np.int8)

        - 'unsigned': smallest unsigned int dtype (min.: np.uint8)

        - 'float': smallest float dtype (min.: np.float32)

    

        As this behaviour is separate from the core conversion to

        numeric values, any errors raised during the downcasting

        will be surfaced regardless of the value of the 'errors' input.

    

        In addition, downcasting will only occur if the size

        of the resulting data's dtype is strictly larger than

        the dtype it is to be cast to, so if none of the dtypes

        checked satisfy that specification, no downcasting will be

        performed on the data.

    

        .. versionadded:: 0.19.0

    

    Returns

    -------

    ret : numeric if parsing succeeded.

        Return type depends on input.  Series if Series, otherwise ndarray

    

    Examples

    --------

    Take separate series and convert to numeric, coercing when told to

    

    >>> import pandas as pd

    >>> s = pd.Series(['1.0', '2', -3])

    >>> pd.to_numeric(s)

    0    1.0

    1    2.0

    2   -3.0

    dtype: float64

    >>> pd.to_numeric(s, downcast='float')

    0    1.0

    1    2.0

    2   -3.0

    dtype: float32

    >>> pd.to_numeric(s, downcast='signed')

    0    1

    1    2

    2   -3

    dtype: int8

    >>> s = pd.Series(['apple', '1.0', '2', -3])

    >>> pd.to_numeric(s, errors='ignore')

    0    apple

    1      1.0

    2        2

    3       -3

    dtype: object

    >>> pd.to_numeric(s, errors='coerce')

    0    NaN

    1    1.0

    2    2.0

    3   -3.0

    dtype: float64

    

    See also

    --------

    pandas.DataFrame.astype : Cast argument to a specified dtype.

    pandas.to_datetime : Convert argument to datetime.

    pandas.to_timedelta : Convert argument to timedelta.

    numpy.ndarray.astype : Cast a numpy array to a specified type.