Arrow community meeting March 15 at 16:00 UTC

[Ian Cook <ia...@ursacomputing.com>]

Hi all,

Our biweekly Arrow community meeting is today at 16:00 UTC / 12:00
EST. For attendees in countries that have not yet switched to Daylight
Saving Time, please note that the time is one hour earlier than usual
in your local time zone.

Zoom meeting URL:
https://zoom.us/j/87649033008?pwd=SitsRHluQStlREM0TjJVYkRibVZsUT09
Meeting ID: 876 4903 3008
Passcode: 958092

The notes for this and future instances of this meeting will be
captured in this Google Doc:
https://docs.google.com/document/d/1xrji8fc6_24TVmKiHJB4ECX1Zy2sy2eRbBjpVJMnPmk/
If you plan to attend the meeting, you are welcome to edit the
document to add the topics that you would like to discuss.

Thanks,
Ian

Re: Arrow community meeting March 15 at 16:00 UTC

[Ian Cook <ia...@ursacomputing.com>]

Below is a summary of the notes from this week's meeting

Attendees:

- Ian Cook
- Raúl Cumplido
- Alenka Frim
- Will Jones
- David Li
- Bryce Mecum
- Rok Mihevc
- Sri Nadukudy
- Weston Pace
- Dane Pitkin
- Matthew Topol
- Joris Van den Bossche

Discussion:

Arrow 12.0.0 Release

- Planned date for code freeze April 11th.
- Plan is for Raúl and Kou to collaborate to manage the release
similar to the 11.0.0 release
- Raúl indicated that under this arrangement, the release process
itself has not been too burdensome, but it has been challenging to
complete all the post-release tasks in a timely manner. For example,
there have been delays updating the conan and vcpkg packages. We could
use help from community members willing to volunteer to perform some
of these post-release tasks
- Please tag any issues that should block the release with:
label:"Priority: Blocker" and milestone:"12.0.0"


Issues we hope to resolve before the code freeze

- Acero ordered execution [1]
- Incorrect hash join results [2]
- Acero scanner work (if time permits)
- Performance regression from row default Parquet group size change [3]
- Slow Parquet reads [4]
- Fixed-shape tensor implementation PR [5]
- Performance regression in writing FixedSizeList to Parquet with
default row group size (if feasible) [6]
- PyArrow bindings to Acero (and removing custom Cython ExecPlan usage) [7]


Removing plasma

- Plasma has been deprecated as of the 10.0.0 release, and its removal
is scheduled to happen in the 12.0.0 release [8] because its original
authors stopped contributing to the Arrow community and forked their
own code for internal use inside another project
- There is ongoing investigation into which alternatives we should
recommend to replace Plasma in different use cases
- The maintainers intend to proceed as per this plan
- Will Jones has sent an email to the dev@ and user@ lists to provide
a reminder about this and seek recommendations about alternatives [9]

[1] https://github.com/apache/arrow/issues/32991
[2] https://github.com/apache/arrow/issues/34474
[3] https://github.com/apache/arrow/issues/34374
[4] https://github.com/apache/arrow/issues/34319
[5] https://github.com/apache/arrow/pull/8510
[6] https://github.com/apache/arrow/issues/34510
[7] https://github.com/apache/arrow/pull/34401
[8] https://lists.apache.org/thread/nw232k2lzmg9kcl8ts475m9ybl34j81p
[9] https://lists.apache.org/thread/1mrx0qg8dflshc4k0fv7g5qm775yr282


On Wed, Mar 15, 2023 at 9:31 AM Ian Cook <ia...@ursacomputing.com> wrote:
>
> Hi all,
>
> Our biweekly Arrow community meeting is today at 16:00 UTC / 12:00
> EST. For attendees in countries that have not yet switched to Daylight
> Saving Time, please note that the time is one hour earlier than usual
> in your local time zone.
>
> Zoom meeting URL:
> https://zoom.us/j/87649033008?pwd=SitsRHluQStlREM0TjJVYkRibVZsUT09
> Meeting ID: 876 4903 3008
> Passcode: 958092
>
> The notes for this and future instances of this meeting will be
> captured in this Google Doc:
> https://docs.google.com/document/d/1xrji8fc6_24TVmKiHJB4ECX1Zy2sy2eRbBjpVJMnPmk/
> If you plan to attend the meeting, you are welcome to edit the
> document to add the topics that you would like to discuss.
>
> Thanks,
> Ian