Hi Jingsong,
Thanks for the clarification, and sorry to misunderstand your first
intention.
What I was talking about is indeed another topic, we can leave it to the
future,
and see if there are any other people who have the same scenarios.
Jingsong Li 于2020年9月3日周四 上午10:56写道:
> Thanks Timo for
Thanks Timo for working on FLIP-107.
Agree, I think it is good.
I'll spend more time to form a FLIP in detail later.
Best,
Jingsong
On Wed, Sep 2, 2020 at 7:12 PM Timo Walther wrote:
> Hi Jingsong,
>
> I haven't looked at your proposal but I think it make sense to have a
> separate FLIP for
Hi Jingsong,
I haven't looked at your proposal but I think it make sense to have a
separate FLIP for the parititioning topic. I'm currently working on an
update to FLIP-107 and would suggest to remove the paritioning topic
there. FLIP-107 will only focus on accessing metadata and expressing
Thanks Konstantin and Benchao for your response.
If we need to push forward the implementation, it should be a FLIP.
My original intention was to unify the partition definitions for batches
and streams:
- What is "PARTITION" on a table? Partitions define the physical storage
form of a table.
Hi Jingsong,
Thanks for bringing up this discussion. I like this idea generally.
I'd like to add some cases we met in our scenarios.
## Source Partition By
There is an use case that users want to do some lookup thing in the UDF,
it's very like the dimension table. It's common for them to cache
Hi Jingsong,
I would like to understand this FLIP (?) a bit better, but I am missing
some background, I believe. So, some basic questions:
1) Does the PARTITION BY clause only have an effect for sink tables
defining how data should be partitioning the sink system or does it also
make a
Hi all,
## Motivation
FLIP-63 [1] introduced initial support for PARTITIONED BY clause to an
extent that let us support Hive's partitioning.
But this partition definition is completely specific to Hive/File
systems, with the continuous development of the system, there are new
requirements:
-