PySpark Scripy and Numpy problem 2

Hi,
this question is the extension of this previus
I cannot Install numpy on company hadoop cluster.
Can I ad my python library via personalization of livy session? ( Node Create Spark Context (Livy)

Can someone help me?

Thank 1000 in advance
Giorgio

Hi @salvatorigio,

You might be able to provide an archive with the dependencies using spark.yarn.dist.archives. Never tested this, it might depend on your Livy configuration if it allows providing them, or needs to be configured in the Livy server config. A blog post about how to create such an archive can be found here: How to Manage Python Dependencies in Spark

Cheers,
Sascha

1 Like

Grazie, ci provo
Giorgio