flink-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Konstantin Knauf <kna...@apache.org>
Subject [ANNOUNCE] Weekly Community Update 2020/16
Date Sun, 19 Apr 2020 21:17:54 GMT
Dear community,

happy to share this (and last) week's community update after a short Easter
break. A lot has happened in the community in the meantime. Stateful
Functions 2.0.0 was released, the releases of Flink 1.10.1 and 1.9.3 are
around the corner, a couple of new FLIPs and blog posts...

...and, of course, Flink Forward 2020 Virtual is on next week Wed - Fri!

Flink Development

* [releases] Apache Flink Stateful Functions 2.0.0 is out. [1] Check out
Stephan's announcement blog post for an overview. [2]

* [releases] Dian Fu recently proposed releasing Flink 1.9.3 [3]. The only
remaining blocker was merged quickly and the first release candidate is out
already [4]

* [releases] There is only one blocker left for Flink 1.10.1 and we
expecting a first release candidate soon. [5]

* [sql] Zhenghua Gao has started a discussion on FLIP-71 to finish initial
end-to-end view support in Flink SQL. [6]

* [sql, hive] Rui Li has authored a FLIP(-123) to increase the
compatibility of Flink with Hive's SQL dialect. It proposes to add an
additional parser for the Hive dialect and to a support a limited set of
DDL and DML features of Hive, that Flink currently does not understand. The
vote has already passed. [7,8]

* [python] Xingbo has started a discussion support Cython for Python User
Defined Functions in the Table API. Quick discussion and vote has already
gone through. [9,10]

* [distribution] Aljoscha has initiated a discussion on releasing a "fat"
and "slim" Flink distribution going forward. The slim distribution would
only contain an absolute minimal set of dependencies (less than today),
while the fat distribution would contain many convenience dependencies like
connectors and formats. The goal is to improve the initial user
experience particularly for Table API/SQL users. Discussion still ongoing.

* [connectors] Dawid proposes (FLIP-124] to add a "Rich" version of
(De)SerializationSchema (adding open/close-method and a Collector). [12]
This will allow to emit an arbitrary number of records for a single source
record, initialization code as well as the registration of metrics in
serializers. Vote has already started. [13,14]

* [docs] Marta proposes to apply to Google's Season of Docs 2020. Season of
Docs brings together technical writers and Open Source projects to improve
their documentation. Marta is looking for (documentation) project ideas as
well as volunteers to mentor potential technical writers. [15]

* [development process] Yun Tang has enabled autolinks from Github commits
to Jirta tickets for all Apache Flink repositories. "FLINK-XXXX" in a
commit message will now automatically link to the corresponding Jira
ticket. Checkout https://github.com/apache/flink/commits/master. [16]

* [development process] Aljoscha reminds everyone to check that their email
address is correctly configured in Github so that it shows up in your
commits to Apache Flink instead of noreply@github.com. [17]

[14] https://flink.apache.org/news/2020/04/07/release-statefun-2.0.0.html

Notable Bugs

[FLINK-16662] [1.10.0] Currently, you can not convert a DataStream of POJOs
to a Table. Fix planned for 1.10.1. [18]

[18] https://issues.apache.org/jira/browse/FLINK-16662

Events, Blog Posts, Misc

* Hequn Chen joined the Apache Flink PMC. Congratulations! [19]

* Seth Wiesman is an Apache Flink Committer now. Congrats! [20]

* Flink Forward San Francisco Virtual will happen next week Wed - Fri. You
can still register & attend for free and listen to over 40 talks by great
speaker.  [21]

* David would like to contribute the material of Ververica's self-paced
Apache Flink training to Apache Flink. [22] The feedback was positive and
the details are discussed in a follow up thread. [23]

* Abdelkrim Hadjidj has published a blog post that implements an imaginary
supply chain use case with an Open Source stream processing stack including
among others tools Apache Flink & Zeppelin. [24]

* Nico started a series of posts on serialization in Apache Flink. A topic
that is often crucial for performance in many DataStream API applications.
His first post explains & compares the different available serializers
available in Apache Flink. [25]

* Also on the Flink blog, Jincheng and Markos recap the latest work on
Python UDF support in the Table API, explain how to get started and have a
look at future work in this area. [26]

[21] https://www.flink-forward.org/sf-2020/conference-program
[26] https://flink.apache.org/2020/04/09/pyflink-udf-support-flink.html


Konstantin (@snntrable)


Konstantin Knauf

View raw message