pandas.DataFrame.to_records

DataFrame.to_records(index=True, convert_datetime64=None)

Convert DataFrame to a NumPy record array.

Index will be put in the ‘index’ field of the record array if requested.

Parameters:

index : boolean, default True

Include index in resulting record array, stored in ‘index’ field.

convert_datetime64 : boolean, default None

Deprecated since version 0.23.0.

Whether to convert the index to datetime.datetime if it is a DatetimeIndex.

Returns:

y : numpy.recarray

See also

DataFrame.from_records

convert structured or record ndarray to DataFrame.

numpy.recarray

ndarray that allows field access using attributes, analogous to typed columns in a spreadsheet.

Examples

>>> df = pd.DataFrame({'A': [1, 2], 'B': [0.5, 0.75]},
...                   index=['a', 'b'])
>>> df
   A     B
a  1  0.50
b  2  0.75
>>> df.to_records()
rec.array([('a', 1, 0.5 ), ('b', 2, 0.75)],
          dtype=[('index', 'O'), ('A', '<i8'), ('B', '<f8')])

The index can be excluded from the record array:

>>> df.to_records(index=False)
rec.array([(1, 0.5 ), (2, 0.75)],
          dtype=[('A', '<i8'), ('B', '<f8')])

By default, timestamps are converted to datetime.datetime:

>>> df.index = pd.date_range('2018-01-01 09:00', periods=2, freq='min')
>>> df
                     A     B
2018-01-01 09:00:00  1  0.50
2018-01-01 09:01:00  2  0.75
>>> df.to_records()
rec.array([(datetime.datetime(2018, 1, 1, 9, 0), 1, 0.5 ),
           (datetime.datetime(2018, 1, 1, 9, 1), 2, 0.75)],
          dtype=[('index', 'O'), ('A', '<i8'), ('B', '<f8')])

The timestamp conversion can be disabled so NumPy’s datetime64 data type is used instead:

>>> df.to_records(convert_datetime64=False)
rec.array([('2018-01-01T09:00:00.000000000', 1, 0.5 ),
           ('2018-01-01T09:01:00.000000000', 2, 0.75)],
          dtype=[('index', '<M8[ns]'), ('A', '<i8'), ('B', '<f8')])