Fix index refcount in unbox_dataframe_column

[Hardcode84]

No description provided.

[AlexanderKalistratov]

@Hardcode84 was you able to check fix suggested by Siu?

[Hardcode84]

@Hardcode84 was you able to check fix suggested by Siu?

Will check tomorrow

[Hardcode84]

Well, when I try to write c.context.nrt.incref(builder, df_typ, dataframe) instead of current solution I'm getting some nonsense like this:

======================================================================
ERROR: test_df_drop_one_column_unboxing (sdc.tests.test_dataframe.TestDataFrame) (index=['a', 'b', 'c', 'd'])
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/ibutygin/src/numba/hpat/sdc/tests/test_dataframe.py", line 1166, in test_df_drop_one_column_unboxing
    pd.testing.assert_frame_equal(sdc_func(df), test_impl(df))
  File "/home/ibutygin/src/numba/numba/numba/dispatcher.py", line 401, in _compile_for_args
    error_rewrite(e, 'typing')
  File "/home/ibutygin/src/numba/numba/numba/dispatcher.py", line 344, in error_rewrite
    reraise(type(e), e, None)
  File "/home/ibutygin/src/numba/numba/numba/six.py", line 668, in reraise
    raise value.with_traceback(tb)
numba.errors.TypingError: Failed in nopython mode pipeline (step: nopython frontend)
Invalid use of Function(<function sdc_pandas_dataframe_drop at 0x7f379eee8dd0>) with argument(s) of type(s): (dataframe((array(float64, 1d, C), array(int64, 1d, C), array(float64, 1d, C)), StringArrayType(), ('A', 'B', 'C'), True), columns=unicode_type)
 * parameterized
In definition 0:
    TypingError: Method drop(). The object columns
 given: unicode_type
 expected: str, tuple of str
    raised from /home/ibutygin/src/numba/hpat/sdc/utilities/sdc_typing_utils.py:75
In definition 1:
    TypingError: Method drop(). The object columns
 given: unicode_type
 expected: str, tuple of str
    raised from /home/ibutygin/src/numba/hpat/sdc/utilities/sdc_typing_utils.py:75
This error is usually caused by passing an argument of a type that is unsupported by the named function.
[1] During: resolving callee type: BoundFunction((<class 'sdc.hiframes.pd_dataframe_type.DataFrameType'>, 'drop') for dataframe((array(float64, 1d, C), array(int64, 1d, C), array(float64, 1d, C)), StringArrayType(), ('A', 'B', 'C'), True))
[2] During: typing of call at /home/ibutygin/src/numba/hpat/sdc/tests/test_dataframe.py (1153)


File "sdc/tests/test_dataframe.py", line 1153:
        def test_impl(df):
            return df.drop(columns='A')
            ^

[AlexanderKalistratov]

Let's merge it as is