Hi Timo and flink-user,
It's been a few weeks and we've made some changes to the application
mentioned on this email. we've also updated for flink 1.4 . We are using
the SQL / Table API with a tumbling window and user defined agg to generate
a SQL query string like:
SELECT measurement, `tag_App
Hi,
I am new to Flink and exploring it. Just curious to know that, do Flink API
support Parquet format like Spark API can easily read and write in Parquet.
--
*Thanks & RegardsMuhammad Imran Tariq*
Hi,
I've been going through various talks on flink's support for queryable
state. Like this talk by Jamie Grier at 2016's Flink forward :
https://www.youtube.com/watch?v=uuv-lnOrD0o
I see how you can easily use this to produce time series data. Eg calculate
the number of events per hour.
But I
Hi aitozi,
`MergingWindowSet` is a Utility, used for keeping track of merging Windows
when using a MergingWindowAssigner in a WindowOperator.
In flink `MergingWindowAssigner` only used for SessionWindow. The
implementations of `MergingWindowAssigner` are `EventTimeSessionWindows`
and `Process
I misuse the key selector. I checked the doc and found it must return
deterministic key, so using random is wrong, but I still could not
understand why it would cause oom.
2017-12-28 21:57 GMT+08:00 Jinhua Luo :
> It's very strange, when I change the key selector to use random key,
> the jvm rep
Yes, that sounds like what Jaxon is looking for. :-) Thanks for the
pointer Eron.
– Ufuk
On Thu, Dec 28, 2017 at 8:13 PM, Eron Wright wrote:
> I believe you can extend the `KeyedDeserializationSchema` that you pass to
> the consumer to check for end-of-stream markers.
>
> https://ci.apache.org/p