arkouda.pandas.conversion

Functions

from_series(...)

Convert a pandas Series to an Arkouda pdarray or Strings.

Module Contents

arkouda.pandas.conversion.from_series(series: pandas.Series, dtype: type | str | None = None) → arkouda.numpy.pdarrayclass.pdarray | arkouda.numpy.strings.Strings

Convert a pandas Series to an Arkouda pdarray or Strings.

If dtype is not provided, the dtype is inferred from the pandas Series (using pandas dtype metadata). If dtype is provided, it is used as an override and normalized via Arkouda’s dtype resolution rules.

In addition to the core numeric and boolean types, this function supports datetime and timedelta Series of any resolution (ns, us, ms, etc.) by converting them to an int64 pdarray of nanoseconds.

Parameters:

• series (pd.Series) – The pandas Series to convert.

• dtype (Optional[Union[type, str]], optional) –

  Optional dtype override. This may be a Python type (e.g. bool), a NumPy scalar type (e.g. np.int64), or a dtype string.

  String-like spellings are normalized to Arkouda string dtype, including "object", "str", "string", "string[python]", and "string[pyarrow]".

Returns:

An Arkouda pdarray for numeric, boolean, datetime, or timedelta inputs, or an Arkouda Strings for string inputs.

Return type:

Union[pdarray, Strings]

Raises:

ValueError – Raised if the dtype cannot be interpreted or is unsupported for conversion.

Examples

>>> import arkouda as ak
>>> import numpy as np
>>> import pandas as pd

Integers:

>>> np.random.seed(1701)
>>> ak.from_series(pd.Series(np.random.randint(0, 10, 5)))
array([4 3 3 5 0])

>>> ak.from_series(pd.Series(['1', '2', '3', '4', '5']), dtype=np.int64)
array([1 2 3 4 5])

Floats:

>>> np.random.seed(1701)
>>> ak.from_series(pd.Series(np.random.uniform(low=0.0, high=1.0, size=3)))
array([0.089433234324597599 0.1153776854774361 0.51874393620990389])

Booleans:

>>> np.random.seed(1864)
>>> ak.from_series(pd.Series(np.random.choice([True, False], size=5)))
array([True True True False False])

Strings (pandas dtype spellings normalized to Arkouda Strings):

>>> ak.from_series(pd.Series(['a', 'b', 'c', 'd', 'e'], dtype="string"))
array(['a', 'b', 'c', 'd', 'e'])

>>> ak.from_series(pd.Series(['a', 'b', 'c'], dtype="string[pyarrow]"))
array(['a', 'b', 'c'])

Datetime (any resolution is accepted and returned as int64 nanoseconds):

>>> ak.from_series(pd.Series(pd.to_datetime(['1/1/2018', np.datetime64('2018-01-01')])))
array([1514764800000000000 1514764800000000000])

Notes

Datetime and timedelta Series are converted to int64 nanoseconds.

String-like pandas dtypes (including object) are treated as string and converted to Arkouda Strings.