[Python] Scalar constructors should be callable

[asfimport]

Currently, to create a Scalar instance you must call the pa.scalar factory function. This works fine for the simplest types, but not for more elaborate ones.

For example, in C++ it is trivial to create a ListScalar from the embedded std::shared_ptr<Array>. But not in Python:

>>> arr = pa.array(["abc", "def", "ghi"])
>>> pa.ListScalar(arr)
Traceback (most recent call last):
  Input In [7] in <cell line: 1>
    pa.ListScalar(arr)
TypeError: __init__() takes exactly 0 positional arguments (1 given)

There is a clumsy workaround that creates an intermediate ListArray with trivial offsets:

>>> list_arr = pa.ListArray.from_arrays([0, len(arr)], arr)
>>> list_arr[0]
<pyarrow.ListScalar: ['abc', 'def', 'ghi']>

Reporter: Antoine Pitrou / @pitrou

Related issues:

• [Python] accept pyarrow values / scalars in constructor functions ? (is related to)

_{Note: This issue was originally created as ARROW-16254. Please see the migration documentation for further details.}

[asfimport]

Antoine Pitrou / @pitrou:
cc @AlenkaF @raulcd

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
For this specific example, the pa.scalar(..) factory function also seems to work?

In [32]: pa.scalar(["abc", "def", "ghi"])
Out[32]: <pyarrow.ListScalar: ['abc', 'def', 'ghi']>

[asfimport]

Antoine Pitrou / @pitrou:
It does not if you pass an existing PyArrow array:

>>> arr = pa.array(["abc", "def", "ghi"])
>>> pa.scalar(arr)
Traceback (most recent call last):
  ...
ArrowInvalid: Could not convert <pyarrow.lib.StringArray object at 0x7f79ab490880>
[
  "abc",
  "def",
  "ghi"
] with type pyarrow.lib.StringArray: did not recognize Python value type when inferring an Arrow data type
>>> pa.scalar(arr, type=pa.list_(arr.type))
Traceback (most recent call last):
  ...
ArrowTypeError: Expected bytes, got a 'pyarrow.lib.StringScalar' object

[asfimport]

Joris Van den Bossche / @jorisvandenbossche:
Ah, yes. Now, that relates to the issue that pa.array(..) also does not accept pyarrow arrays as scalars, see ARROW-5295

[github-actions]

This issue has been marked as stale because it has had no activity in the past 365 days. Please remove the stale label or comment below, or this issue will be closed in 14 days. If this improvement is still desired but has no current owner, please add the 'Status: needs champion' label.