On Tue, Apr 7, 2020, 12:04 PM Antoine Pitrou <anto...@python.org> wrote:
>
> Le 07/04/2020 à 18:49, Todd Lipcon a écrit :
> >>
> >> Hmm, the spec may not be clear enough on this, but if you move a child
> >> and release the parent, then the other children are not usable anymore.
> >>
> >> In your case, you don't call release() on every child. You just call
> >> release() on the parent once you are done with all children. So the
> >> synchronization has to be managed on the consumer side, the producer
> >> doesn't see any of it.
> >>
> >
> > I'm not sure I follow. If the consumer takes a top-level ArrowArray, and
> > then moves out each of the children, then calls release() on the
> top-level
> > array, the expectation is that the chlidren stay valid, according to the
> > spec.
>
> According to the spec, once you move a child, you are supposed to
> release the parent ("immediately afterwards"). Then by construction you
> cannot move other children. I don't know if we want to make it possible
> to move several children before releasing the parent. Abstractly, that
> may sound desirable, but we should check that doesn't impose further
> constraints on the producer.
>
Re-orienting the discussion on something more concrete, suppose that an
ArrowArray is used to convey a result set from a database query, and
suppose that the resources associated with each column in the result set
are independent of the other columns. Should it not be possible to select a
subset of the columns and release the resources of the columns that are not
needed?
> However, if as a consumer you're simply processing all children, then
> there's no need to move them. Just release the parent when you are
> finished with all children.
>
> Regards
>
> Antoine.
>
