Re: How to make a generic key for groupBy

Hi! I have some ideas, let me see if I can make them concrete until tomorrow... Greetings, Stephan On Mon, Apr 27, 2015 at 5:29 PM, LINZ, Arnaud wrote: > Hi, > > I see. My Key class is an abstract class, which subclasses are Key1, > Key2 etc, so it’s very like a tuple. It is heavily used in

Re: Flink Java 8 problem (no lambda, simple code)

The effect of disabling the closure cleaner is that some cases can throw you a "NonSerializableException". If you do not see that, there is no problem anyways, no unexpected side effect :-) Have you tried the latest version of the Flink master? We updated the code (2 days ago) to use as newer ver

RE: Flink Java 8 problem (no lambda, simple code)

Hi, With cluster.getConfig().disableClosureCleaner() it works, thanks ; what are the sides effect of this workaround? My code is just a plain Junit test, the only code that is called outside of what I gave is : public static String chercher(String nom) { String retour = null;

Re: error when eun program left outer join

I solve mu problem very thanks -- View this message in context: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/error-when-eun-program-left-outer-join-tp1141p1146.html Sent from the Apache Flink User Mailing List archive. mailing list archive at Nabble.com.

RE: How to make a generic key for groupBy

Hi, I see. My Key class is an abstract class, which subclasses are Key1, Key2 etc, so it’s very like a tuple. It is heavily used in “non-distributed” hash maps once the dataset is reduced to fit on a single JVM. It exposes the common contract that I need (such as getHeadKey(), getLastl(), or mak

Re: Difference between using a global variable and broadcasting a variable

Hi! I put a quick summary into the wiki. For future reference. https://cwiki.apache.org/confluence/display/FLINK/Variables+Closures+vs.+Broadcast+Variables Greetings, Stephan On Mon, Apr 27, 2015 at 11:10 AM, Stephan Ewen wrote: > Adding to Fabian's and Sebastian's answer: > > > Variable in

Re: error when eun program left outer join

implement left outer join from two dataset Customer and Orders using Tuple data type -- View this message in context: http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/error-when-eun-program-left-outer-join-tp1141p1143.html Sent from the Apache Flink User Mailing List archive.

Re: error when eun program left outer join

Hi, what data are you using? The exception says "NullFieldException: Field 1 is null, but expected to hold a value.". Maybe the data is not in the right format? On Mon, Apr 27, 2015 at 2:32 PM, hagersaleh wrote: > I want implement left outer join from two dataset i use Tuple data type > > > pa

error when eun program left outer join

I want implement left outer join from two dataset i use Tuple data type package org.apache.flink.examples.java.relational; import org.apache.flink.api.common.functions.CoGroupFunction; import org.apache.flink.api.common.functions.MapFunction; import org.apache.flink.api.java.DataSet; import org

Re: Difference between using a global variable and broadcasting a variable

Adding to Fabian's and Sebastian's answer: Variable in Closure (global variable) -- - Happens when you reference some variable in the program from a function. The variable becomes part of the Function's closure. - The variable is distributed w

Re: java.io.EOFException: Premature EOF: no length prefix available

To add to Robert's answer: The two parameters "taskmanager.heartbeat-interval: 1" and " jobmanager.max-heartbeat-delay-before-failure.sec: 90" are deprecated in the current 0.9 master. We need to use the akka parameters for the heartbeat in 0.9. Example: akka.watch.heartbeat.interval: 10s a

Re: Flink - Avro - AvroTypeInfo issue - Index out of bounds exception

Hi Filip! We pushed a fix to the master end of last week, which addressed an issue with pre-mature buffer pool deallocation. Can you try if the latest version fixes your problem? Greetings, Stephan On Sun, Apr 26, 2015 at 5:16 PM, Filip Łęczycki wrote: > Hi Stephan, > > You are right, sorry

Re: Difference between using a global variable and broadcasting a variable

You should also be aware that the value of a static variable is only accessible within the same JVM. Flink is a distributed system and runs in multiple JVMs. So if you set a value in one JVM it is not visible in another JVM (on a different node). In general, I would avoid to use static variables i

Re: Objects deserialization on Jobmanager

Hi Ventura! I hope you can get along without editing the Flink runtime, with the help of overriding the open/close methods of the RichFunctions and do your initialization there. They are called once per life of a parallel function (except in iterations, where they are called once per superstep, bu

Re: java.io.EOFException: Premature EOF: no length prefix available

Hi, it looks like there the TaskManager on 172.16.20.112:41265 was suffering from a network disruption (because the HDFS connection seems to be affected as well). Are the clocks on the JobManager and the TaskManager out of sync? Because the jobmanager is complaining on 07 about the missing TM, bu

Re: flink ml - k-means

Hi Alexander and Till, thanks for your informations, I look forward to the release. I'm curious how well is flink ml against mahout und spark ml. best regerds Paul 2015-04-27 9:23 GMT+02:00 Till Rohrmann : > Hi Paul, > > if you can't wait, a vanilla implementation is already contained as part >

Re: flink ml - k-means

Hi Paul, if you can't wait, a vanilla implementation is already contained as part of the Flink examples. You should find it under flink/flink-examples. But we will try to add more clustering algorithms in the near future. Cheers, Till On Apr 26, 2015 11:14 PM, "Alexander Alexandrov" < alexander.