First of all I have been looking around for this problem a while now, and I can see there exist other solutions regarding this, however nothing for the Apache Spark version 2.3.1.
To be short, I am trying to create an application that uses bahir to perform analytics twitter messages in spark.
However, I am using Apache Spark version 2.3.1, so I found 2.3.0-SNAPSHOT
But when I try to use this bin/spark-shell --packages org.apache.bahir:spark-streaming-twitter_2.11:2.3.0-SNAPSHOT
, it can't be found, from my local spark-shell:
:: USE VERBOSE OR DEBUG MESSAGE LEVEL FOR MORE DETAILS Exception in thread "main" java.lang.RuntimeException: [unresolved dependency: org.apache.bahir#spark-streaming-twitter_2.11;2.3.0-SNAPSHOT: not found]
I may be a stupid assumption but I figured maybe that the 2.3.0 would work for 2.3.1.
I can fetch the 2.2.1 version and the twitter4j seems to work, however I still get problems with the actual streaming.twitter_
to work with Spark 2.3.1. From my local spark-shell:
scala> import org.apache.spark.streaming.twitter._ <console>:23: error: object twitter is not a member of package org.apache.spark.streaming import org.apache.spark.streaming.twitter._
^
Would be nice to know if somebody knows if it is usable with Apache Spark 2.3.1?
Or am I just simply forced to downgrade my Spark version to make it work?
I am doing this in a notebook called Zeppelin but I have also tried to do this outside of Zeppelin so seems not to have anything to do with the notebook I am using.
Thankful for any insights.
I faced the same issue. I can't lower the spark, because I need 2.3 with another helium.. So I'm going to try bahir's dependency
I will let you know, if it works...
Edit: It did work.
It solves issue with non-existing new twitter libraries and also it solves the issue with missing classes from "old spark" in "new spark".
I was able to run sum examples using javascript leaflet map and spark 2.3 sparkstreaming :)