`open_ipc_buffer` sounds fine to me. > Btw, py_buffer(<Python bytes object>) makes a copy, so it is not a pure > logical buffer creator.
I don't think so: ```python >>> b = b"12345" * 50 >>> buf = pa.py_buffer(b) >>> buf.address - id(b) 32 >>> id(b) < buf.address < id(b) + sys.getsizeof(b) True ``` The Arrow buffer does point inside the bytes object (just at the start of its payload, which is allocated contiguously to the 4-words header). [ Full content available at: https://github.com/apache/arrow/pull/2536 ] This message was relayed via gitbox.apache.org for [email protected]
