+1 (non-binding)
On Tue, Feb 16, 2021 at 11:11 PM Maxim Gekk
wrote:
> +1 (non-binding)
>
> On Wed, Feb 17, 2021 at 9:54 AM Wenchen Fan wrote:
>
>> +1
>>
>> On Wed, Feb 17, 2021 at 1:43 PM Dongjoon Hyun
>> wrote:
>>
>>> +1
>>>
>>> Bests,
>>> Dongjoon.
>>>
>>>
>>> On Tue, Feb 16, 2021 at 2:27 AM
I did a simple benchmark (adding two long values) to compare the
performance between
1. native expression
2. the current UDF
3. new UDF with individual parameters
4. new UDF with a row parameter (with the row object cached)
5. invoke a static method (to explore the possibility of speeding up
statel
Thank you so much for sharing the progress, Wenchen! Also, thank you,
Hyukjin.
Bests,
Dongjoon.
On Wed, Feb 17, 2021 at 2:49 AM Wenchen Fan wrote:
> I did a simple benchmark (adding two long values) to compare the
> performance between
> 1. native expression
> 2. the current UDF
> 3. new UDF wi
Thanks, Hyukjin. I think that's a fair summary. And I agree with the idea
that we should focus on what Spark will do by default.
I think we should focus on the proposal, for two reasons: first, there is a
straightforward path to incorporate Wenchen's suggestion via
`SupportsInvoke`, and second, th
+1
I've looked around the jira tickets and I couldn't find any blocker in the
SQL part.
Also, I ran the tests on aws env and I couldn't find any critical error
there, too.
On Wed, Feb 17, 2021 at 5:21 PM John Zhuge wrote:
> +1 (non-binding)
>
> On Tue, Feb 16, 2021 at 11:11 PM Maxim Gekk
> wr
I think I'm +1 on this, in that I don't see any more test failures than I
usually do, and I think they're due to my local env, but is anyone seeing
these failures?
- includes jars passed in through --jars *** FAILED ***
Process returned with exit code 1. See the log4j logs for more detail.
(Spark
I didn't see them. Could you describe your environment: OS, Java,
Maven/SBT, profiles?
On Wed, Feb 17, 2021 at 6:26 PM Sean Owen wrote:
> I think I'm +1 on this, in that I don't see any more test failures than I
> usually do, and I think they're due to my local env, but is anyone seeing
> these
I'm on Ubuntu 20, Java 8, Maven, with most every profile enabled (Hive,
YARN, Mesos, K8S, SparkR, etc). I think it's probably transient or specific
to my env; just checking if anyone else sees this. Obviously the main test
builds do not fail on Jenkins.
On Wed, Feb 17, 2021 at 10:47 PM Dongjoon Hy
I don't see any objections to the rest of the proposal (loading functions
from the catalog, function binding stuff, etc.) and I assume everyone is OK
with it. We can commit that part first.
Currently, the discussion focuses on the `ScalarFunction` API, where I
think it's better to directly take th
