Hi - this seems to be an issue with the way the python code is imported from a
jar or from a Spark package. I ran into the same. I tried bt couldn't find any
guideline on how a Spark package should make its Python binding available. If
you would open an issue at graph frame, I could chime in there.
_____________________________
From: enzo <e...@smartinsightsfromdata.com>
Sent: Friday, March 4, 2016 4:31 PM
Subject: graphframes: errors adding dependencies with pyspark
To: <users@zeppelin.incubator.apache.org>
Experimenting with new package graphframes for Spark. Kindly tell me
what do I do wrong.
Here is my code:
%dep // sort dependencies
z.load("graphframes:graphframes:0.1.0-spark1.6")
This gives (so annoying the reference to %dep been deprecated!):
DepInterpreter(%dep) deprecated. Load dependency through GUI
interpreter menu instead. res0: org.apache.zeppelin.dep.Dependency
= org.apache.zeppelin.dep.Dependency@7bf2fe37
%pyspark from graphframes import *
I get:
Traceback (most recent call last): File
"/var/folders/m4/yfyjwcpj6rv5nq730bl9vr1h0000gp/T//zeppelin_pyspark.py", line
232, in <module> eval(compiledCode) File "<string>",
line 1, in <module> ImportError: No module named 'graphframes'
What do I do wrong?
Thanks
Enzo e...@smartinsightsfromdata.com
