[C++] Refactor temporal casts to work with Scalar inputs

[asfimport]

Currently, the casting for time-like data is done with the ShiftTime function. It might be possible to simplify this with ArrayDataVisitor (to avoid looping / checking the bitmap).

Reporter: Joris Van den Bossche / @jorisvandenbossche

Related issues:

• [C++] C++ array kernels framework and execution buildout (umbrella issue) (is a child of)

_{Note: This issue was originally created as ARROW-6974. Please see the migration documentation for further details.}

[asfimport]

Wes McKinney / @wesm:
I edited the title to reframe the issue in the context of the new framework

[asfimport]

Antoine Pitrou / @pitrou:
This works now:

>>> s = pa.scalar(datetime.now())
>>> s
<pyarrow.TimestampScalar: datetime.datetime(2021, 11, 23, 19, 48, 19, 732990)>
>>> pc.cast(s, pa.timestamp('ns'))
<pyarrow.TimestampScalar: Timestamp('2021-11-23 19:48:19.732990')>
>>> pc.cast(s, pa.timestamp('ns', 'Europe/Paris'))
<pyarrow.TimestampScalar: Timestamp('2021-11-23 20:48:19.732990+0100', tz='Europe/Paris')>
>>> pc.cast(s, pa.timestamp('ns', 'UTC'))
<pyarrow.TimestampScalar: Timestamp('2021-11-23 19:48:19.732990+0000', tz='UTC')>