You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@zeppelin.apache.org by Renato Perini <re...@gmail.com> on 2015/10/01 01:48:17 UTC
Re: Zeppelin first time user: some questions.
Sorry, I meant "with Zeppelin".
Basically I want to write Java Spark Code in the Zeppelin notebook.
Thank you.
Il 28/09/2015 09:05, Vinay Shukla ha scritto:
> Spark supports Scala, Java, Python & R.
> Zeppelin supports the first 3, R support is in progress.
>
> On Monday, September 28, 2015, Eugene <blackorangebox@gmail.com
> <ma...@gmail.com>> wrote:
>
> Regarding Q2, it's not really possible currently.
>
> See this feature request:
> https://issues.apache.org/jira/browse/ZEPPELIN-320
>
> 2015-09-28 4:08 GMT+04:00 Renato Perini <renato.perini@gmail.com
> <javascript:_e(%7B%7D,'cvml','renato.perini@gmail.com');>>:
>
> Hello!
> I want to evaluate Apache Zeppelin for some experiments in
> analytics at my company.
> Currently we have a custom setup for Apache Spark on Amazon
> EC2. This small cluster is used for two purposes:
> 1) Routing logs coming from an Apache Flume instance and
> writing them to an Apache Cassandra database through Spark
> Streaming.
> 2) Analyzing written logs on Cassandra to produce some
> (simple) stats for the data stored on Cassandra (producing
> ad-hoc tables with processed results).
>
> I have been asked to produce a simple web app demo to show the
> results of the computations stored on Cassandra and I have
> thought Zeppelin can facilitate my homeworks.
> Now I have some questions, before I even attempt to install it:
>
> 1) Can I use Java as the programming language for interacting
> with Apache Spark? Or I'm constrained with Scala / Python?
> 2) Can I export images, graphs, etc. to directly use them into
> a separate web app? Would it be difficult? Zeppelin can update
> those images interactively as data updates on the tables and
> without user intervention?
> 3) Can I read data directly from Cassandra without any further
> processing? I have setup some tables that contains already
> processed data, so the idea is to read those tables directly
> for showing some graphs (maybe using loading data through
> SparkSQL).
>
> Thank you and keep up the good work.
>
> Renato Perini.
>
>
>
>
> --
>
>
> Best regards,
> Eugene.
>