Hi Umesh,
hoodie-spark-bundle is a shaded fat jar whereas hoodie-spark is not a fat jar.
Balaji.V
On Thursday, March 7, 2019, 8:50:22 AM PST, Umesh Kacha
<[email protected]> wrote:
Ok Balaji thanks will try to use hoodie-spark-bundle show below. What is
the different between two though hoodie-spark and hoodie-spark-bundle?
<groupId>com.uber.hoodie</groupId>
<artifactId>hoodie-spark-bundle</artifactId>
<version>0.4.5</version>
On Thu, Mar 7, 2019 at 9:14 PM Balaji Varadarajan
<[email protected]> wrote:
> Hi Umesh,
> From your email, it looks like you are using hoodie-spark jar. Please use
> hoodie-spark-bundle jar instead of hoodie-spark jar. Here is an example
> from quickstart:
> Quickstart | Hudi
>
>
> |
> |
> | |
> Quickstart | Hudi
>
>
> |
>
> |
>
> |
>
>
>
>
> Balaji.V
> On Thursday, March 7, 2019, 7:00:17 AM PST, Vinoth Chandar <
> [email protected]> wrote:
>
> Hi Umesh,
>
> Normally spark users build a fat/uber jar with all depedencies. This is the
> same process to be followed for hoodie-spark as well, which will pull all
> the dependencies like hoodie-common as well.
>
> Are you doing something similar?
>
> Thanks
> Vinoth
>
> On Thu, Mar 7, 2019 at 2:11 AM Umesh Kacha <[email protected]> wrote:
>
> > Hi I am trying to use Hoodie 0.4.4 jar in spark-shell and when I try to
> > save spark DataFrame as hoodi table I get the exception saying
> >
> > NoClassDefFoundError: com/uber/hoodie/exception/HoodiException
> >
> > Similar error occurs for HoodieWriteConfig values and
> > DataSourceWriteOptions classes.
> >
> > Do I need to include others jars I can see hoodie-spark-0.4.4.jar is uber
> > fat jar which contains all transitive depenendices? No?
> >
> > Please advise. Thank you.
> >
>
