[ARROW-2853] [Python] Implementing support for zero copy NumPy arrays in libarrow_python

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Duplicate
• Component/s: C++, Python
• External issue URL: https://github.com/apache/arrow/issues/19227
• Assignee: Unassigned
• Reporter: Florian Rathgeber
• Created: 15/Jul/18 10:21
• Updated: 11/Jan/23 07:23
• Resolved: 17/Feb/21 16:30

Description

Implementing support for zero copy NumPy arrays in libarrow_python (i.e. in C++). We can utilize common code paths with `to_pandas` and toggle between NumPy-for-pandas and NumPy-for-NumPy behavior (and use the `zero_copy_only` flag where needed).

Issue Links

• is superceded by: ARROW-6749 [Python] Conversion of non-ns timestamp array to numpy gives wrong values
• relates to: ARROW-564 [Python] Add methods to return vanilla NumPy arrays (plus boolean mask array if there are nulls)