You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@beam.apache.org by Brian Hulette <bh...@google.com> on 2022/06/29 00:58:37 UTC
Minor(?) regression in Python 2.40.0
Hi everyone,
We just detected what looks like a minor regression in Python which made it
into the 2.40.0 release. I wrote up the details in an issue [1].
tl;dr: Expanding a beam.FlatMap constructed with a built-in function (e.g.
sum) will raise an error.
I think the blast radius on this is actually quite small. It's hard to
devise a use-case where a FlatMap constructed with a built-in function
would work anyway, since the function must return an iterable. So I'm
planning to just document the known issue and send a fix to master. If
anyone disagrees with this assessment please let me know.
Thanks,
Brian
[1] https://github.com/apache/beam/issues/22091
Re: Minor(?) regression in Python 2.40.0
Posted by Ahmet Altay <al...@google.com>.
Thank you!
On Wed, Jun 29, 2022 at 1:04 PM Brian Hulette <bh...@google.com> wrote:
> Blog post updated [1] and fix is out for review [2].
>
> [1] https://github.com/apache/beam/pull/22101
> [2] https://github.com/apache/beam/pull/22104
>
> On Tue, Jun 28, 2022 at 8:42 PM Ahmet Altay <al...@google.com> wrote:
>
>> Thank you. Could you update the Beam 2.40.0 blog post (
>> https://beam.apache.org/blog/beam-2.40.0/) to add this as a known issue?
>>
>> On Tue, Jun 28, 2022 at 5:59 PM Brian Hulette <bh...@google.com>
>> wrote:
>>
>>> Hi everyone,
>>>
>>> We just detected what looks like a minor regression in Python which made
>>> it into the 2.40.0 release. I wrote up the details in an issue [1].
>>>
>>> tl;dr: Expanding a beam.FlatMap constructed with a built-in function
>>> (e.g. sum) will raise an error.
>>>
>>> I think the blast radius on this is actually quite small. It's hard to
>>> devise a use-case where a FlatMap constructed with a built-in function
>>> would work anyway, since the function must return an iterable. So I'm
>>> planning to just document the known issue and send a fix to master. If
>>> anyone disagrees with this assessment please let me know.
>>>
>>> Thanks,
>>> Brian
>>>
>>> [1] https://github.com/apache/beam/issues/22091
>>>
>>
Re: Minor(?) regression in Python 2.40.0
Posted by Brian Hulette <bh...@google.com>.
Blog post updated [1] and fix is out for review [2].
[1] https://github.com/apache/beam/pull/22101
[2] https://github.com/apache/beam/pull/22104
On Tue, Jun 28, 2022 at 8:42 PM Ahmet Altay <al...@google.com> wrote:
> Thank you. Could you update the Beam 2.40.0 blog post (
> https://beam.apache.org/blog/beam-2.40.0/) to add this as a known issue?
>
> On Tue, Jun 28, 2022 at 5:59 PM Brian Hulette <bh...@google.com> wrote:
>
>> Hi everyone,
>>
>> We just detected what looks like a minor regression in Python which made
>> it into the 2.40.0 release. I wrote up the details in an issue [1].
>>
>> tl;dr: Expanding a beam.FlatMap constructed with a built-in function
>> (e.g. sum) will raise an error.
>>
>> I think the blast radius on this is actually quite small. It's hard to
>> devise a use-case where a FlatMap constructed with a built-in function
>> would work anyway, since the function must return an iterable. So I'm
>> planning to just document the known issue and send a fix to master. If
>> anyone disagrees with this assessment please let me know.
>>
>> Thanks,
>> Brian
>>
>> [1] https://github.com/apache/beam/issues/22091
>>
>
Re: Minor(?) regression in Python 2.40.0
Posted by Ahmet Altay <al...@google.com>.
Thank you. Could you update the Beam 2.40.0 blog post (
https://beam.apache.org/blog/beam-2.40.0/) to add this as a known issue?
On Tue, Jun 28, 2022 at 5:59 PM Brian Hulette <bh...@google.com> wrote:
> Hi everyone,
>
> We just detected what looks like a minor regression in Python which made
> it into the 2.40.0 release. I wrote up the details in an issue [1].
>
> tl;dr: Expanding a beam.FlatMap constructed with a built-in function (e.g.
> sum) will raise an error.
>
> I think the blast radius on this is actually quite small. It's hard to
> devise a use-case where a FlatMap constructed with a built-in function
> would work anyway, since the function must return an iterable. So I'm
> planning to just document the known issue and send a fix to master. If
> anyone disagrees with this assessment please let me know.
>
> Thanks,
> Brian
>
> [1] https://github.com/apache/beam/issues/22091
>