Sorry I don't fully understand your saying. I think you can use Scala/Java API to do data transformation and register as a table, and then use Python API to do ML pipeline, not sure what is your specific scenario.
Thanks Jerry 2018-02-06 18:09 GMT+08:00 simsicon <[email protected]>: > Thanks Jerry, please let me give an example like, if I am going to submit > an application that using Scala API, but I would like to prepend a > customization transformation function using Python at the beginning of the > application, what should I do exactly? And is SparkSession in SparkContext > > <https://livy.incubator.apache.org/docs/latest/api/scala/index.html#org.apache.livy.scalaapi.ScalaJobContext>refers > to the same concept of session in REST api? > > On Tue, Feb 6, 2018 at 3:22 PM, Saisai Shao <[email protected]> > wrote: > >> IIUC, I assume you're using JobAPI rather than REST APIs. Yes, you can >> create a session using Java/Scala job API, and then manipulate with Python >> APIs. >> >> Thanks >> Jerry >> >> 2018-02-06 13:43 GMT+08:00 simsicon <[email protected]>: >> >>> Hi, >>> >>> I am pretty new to livy, and I am looking into the livy to do some >>> research if livy is capable as infrastructure of a data mining platform, >>> which means we may want to pipe different language operators together, so I >>> am excited to see the multiple languages support in the latest release >>> 0.5.0. >>> >>> I wonder if we can use scala api to create a session, then post >>> different language statement to that session? Does this make any sense? >>> >>> Thanks >>> Z >>> >>> -- >>> >>> Thanks&Best Regards >>> >>> Zhang Zheng >>> >> >> > > > -- > > Thanks&Best Regards > > Zhang Zheng >
