Thanks for the inputs, Aljoscha & Till.
# Weight Configuration
I think exposing the knobs incrementally is a good idea. However, I'm not
sure about non-configurable as the first step.
Currently, users can tune memory for rocksdb
('taskmanager.memory.managed.size') and python
('python.fn-execu
Jingsong Lee created FLINK-19121:
Summary: Avoid access HDFS in HiveBulkWriterFactory
Key: FLINK-19121
URL: https://issues.apache.org/jira/browse/FLINK-19121
Project: Flink
Issue Type: New Fe
> a Row has two modes represented by an internal boolean flag
`hasFieldOrder`
+1 confusion with Dawid that what's the result when index-based setters and
name-based setters are mixed used.
And name-based setters look like append instead of set.
It reminds me of Avro's `GenericRecord`, We should s
Timo, Thanks for the discussion
I have only read the "Conversion of DataStream to Table" part so i would only
put some objections there ~
> StreamTableEnvironment.fromInsertStream(DataStream): Table
At first glance, from the perspective of a user, i'm confused by why we must
dintinguish on the
Jingsong raised a good point. We need to be more careful when deprecating
APIs.
For example, tEnv#createTemporaryView was introduced in release-1.10, users
became familiar with this API in the previous release, but now we want to
deprecate it in the next release.
I also have some concerns about de
Thanks Timo for driving.
My first impression is, can we not deprecate these API?
- StreamTableEnvironment.fromDataStream(DataStream): Table
- StreamTableEnvironment.fromDataStream(DataStream, Expression...):
Table
- StreamTableEnvironment.createTemporaryView(String, DataStream,
Expression...): Uni
Tang Yan created FLINK-19120:
Summary: Failed to build Flink with HBase2.1 even using the latest
Flink release version.
Key: FLINK-19120
URL: https://issues.apache.org/jira/browse/FLINK-19120
Project: Fli
Dian Fu created FLINK-19119:
---
Summary: Update the documentation to use Expression instead of
strings in the Python Table API
Key: FLINK-19119
URL: https://issues.apache.org/jira/browse/FLINK-19119
Project:
Dian Fu created FLINK-19118:
---
Summary: Support Expression in the operations of the Python Table
API
Key: FLINK-19118
URL: https://issues.apache.org/jira/browse/FLINK-19118
Project: Flink
Issue Typ
Xintong Song created FLINK-19117:
Summary: FLINK-18620
Key: FLINK-19117
URL: https://issues.apache.org/jira/browse/FLINK-19117
Project: Flink
Issue Type: Improvement
Reporter: Xin
McClone created FLINK-19115:
---
Summary: Null value fields cannot be ignored in ElasticSearch Sink
Key: FLINK-19115
URL: https://issues.apache.org/jira/browse/FLINK-19115
Project: Flink
Issue Type: B
Dian Fu created FLINK-19116:
---
Summary: Support more kinds of data for expressions.lit in the
Python Table API
Key: FLINK-19116
URL: https://issues.apache.org/jira/browse/FLINK-19116
Project: Flink
Dian Fu created FLINK-19114:
---
Summary: Introduce Expression class in the Python Table API
Key: FLINK-19114
URL: https://issues.apache.org/jira/browse/FLINK-19114
Project: Flink
Issue Type: Sub-task
Roman Khachatryan created FLINK-19113:
-
Summary: Add support for checkpointing with selectable inputs
Key: FLINK-19113
URL: https://issues.apache.org/jira/browse/FLINK-19113
Project: Flink
Nico Kruber created FLINK-19112:
---
Summary: No access to metric group in ScalarFunction when
optimizing
Key: FLINK-19112
URL: https://issues.apache.org/jira/browse/FLINK-19112
Project: Flink
Is
Thanks for creating this FLIP Xintong.
I agree with the previous comments that the memory configuration should be
as easy as possible. Every new knob has the potential to confuse users
and/or allows him to shoot himself in the foot. Consequently, I am +1 for
the first proposal in the FLIP since it
Thanks for the healthy discussion Jark and Dawid.
6. "Row#setFieldNames(@Nullable Map fieldNames)"
Yes, I'm concerned about about the per-record performance. A converter
or serializer should prepare an immutable Map instance before (stored in
a member variable) that is simply passed to every n
Tobias Kaymak created FLINK-19111:
-
Summary: Flink Docker image creates checkpoints as root user and
hits permission denied afterwards
Key: FLINK-19111
URL: https://issues.apache.org/jira/browse/FLINK-19111
Hi Timo,
Thanks for the quick response.
5. "StreamStatementSet#attachToStream()"
Joining or using connect() with a different DataStream is a good case.
cc @Godfrey , what do you think about the `attachToStream()` API?
6. "Row#setFieldNames(@Nullable Map fieldNames)"
> We need a Map for constant
Hi all,
I really like the ideas of this FLIP. I think it improves user
experience quite a bit. I wanted to add just two comments:
1. As for the StatementSet I like the approach described in the FLIP for
its simplicity. Moreover the way I see it is that if a user wants to
work with DataStream, the
Hi,
playing devils advocate here: should we even make the memory weights
configurable? We could go with weights that should make sense for most
cases in the first version and only introduce configurable weights when
(if) users need them.
Regarding where/how things are configured, I think tha
Wei Zhong created FLINK-19110:
-
Summary: Flatten current PyFlink documentation structure
Key: FLINK-19110
URL: https://issues.apache.org/jira/browse/FLINK-19110
Project: Flink
Issue Type: Sub-tas
Hmm, it seems I left out the Dev ML in my mail. Looping that back in..
On 28.08.20 13:54, Dawid Wysakowicz wrote:
@Aljoscha Let me bring back to the ML some of the points we discussed
offline.
Ad. 1 Yes I agree it's not just about scheduling. It includes more
changes to the runtime. We might n
Hi Jark,
thanks for the detailed review. Let me answer your concerns:
## Conversion of DataStream to Table
1. "We limit the usage of `system_rowtime()/system_proctime` to the leaf
of a QueryOperation tree in the validation phase."
I'm fine with allowing `system_proctime` everywhere in the quer
Thanks for letting the community know Roman!
Cheers,
Till
On Tue, Sep 1, 2020 at 10:57 AM Khachatryan Roman <
khachatryan.ro...@gmail.com> wrote:
> Hello,
>
> Yesterday the machine executing Flink benchmarks was replaced due to
> hardware problems.
> The HW configuration is different, so the res
Hi Alexey,
Glad to hear that your are interested the K8s HA support.
Roman's answer is just on point.
"FileSystemBlobStore" is trying to store the user jars, job graph, etc. on
the distributed storage(e.g. HDFS, S3, GFS). So when the
JobManager failover, it could fetch the blob data from remote
Hello,
Yesterday the machine executing Flink benchmarks was replaced due to
hardware problems.
The HW configuration is different, so the results may differ from what we
had previously.
Regards,
Roman
@Seth - Possibly, it turns out to be more tricky than originally thought,
we will have it as a stretch goal for this release.
On Mon, Aug 31, 2020 at 3:12 PM Seth Wiesman wrote:
> +1 for Sept 10. Do you think we'd be able to get a fix for FLINK-18894 by
> then?
>
> https://issues.apache.org/jira
David Anderson created FLINK-19109:
--
Summary: Split Reader eats chained periodic watermarks
Key: FLINK-19109
URL: https://issues.apache.org/jira/browse/FLINK-19109
Project: Flink
Issue Type:
Danny Chen created FLINK-19108:
--
Summary: Stop expanding the identifiers with scope aliased by the
system with 'EXPR$' prefix
Key: FLINK-19108
URL: https://issues.apache.org/jira/browse/FLINK-19108
Proje
30 matches
Mail list logo
