[ARROW-4111] [Python] Create time types from Python sequences of integers

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.15.0
• Component/s: Python
• Labels: beginner
• External issue URL: https://github.com/apache/arrow/issues/20702
• Assignee: Wes McKinney
• Reporter: Wes McKinney
• Created: 23/Dec/18 21:33
• Updated: 11/Jan/23 07:31
• Resolved: 23/Aug/19 14:27

Description

This works for dates, but not times:

>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> traceback >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

    def test_to_pandas_deduplicate_date_time():
        nunique = 100
        repeats = 10
    
        unique_values = list(range(nunique))
    
        cases = [
            # array type, to_pandas options
            ('date32', {'date_as_object': True}),
            ('date64', {'date_as_object': True}),
            ('time32[ms]', {}),
            ('time64[us]', {})
        ]
    
        for array_type, pandas_options in cases:
>           arr = pa.array(unique_values * repeats, type=array_type)

pyarrow/tests/test_convert_pandas.py:2392: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
pyarrow/array.pxi:175: in pyarrow.lib.array
    return _sequence_to_array(obj, mask, size, type, pool, from_pandas)
pyarrow/array.pxi:36: in pyarrow.lib._sequence_to_array
    check_status(ConvertPySequence(sequence, mask, options, &out))
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

>   raise ArrowInvalid(message)
E   pyarrow.lib.ArrowInvalid: ../src/arrow/python/python_to_arrow.cc:1012 : ../src/arrow/python/iterators.h:70 : Could not convert 0 with type int: converting to time32

Issue Links

• relates to: ARROW-6227 [Python] pyarrow.array() shouldn't coerce np.nan to string