Hi, if we want to use the <metaflow-pyspark> plugi...
# ask-metaflow
w
Hi, if we want to use the metaflow-pyspark plugin, does it need to be in declared in the flow dependencies (ie
@conda
or
@pip_base
)?
1
v
no, extensions get included automatically
are you going to use Spark on EMR Serverless?
w
A colleague is going to try it out. Are there dragons?
v
EMR Serverless should work out of the box. Elsewhere be dragons 🐉
(it's just that spark config works differently in different environments - we can add support as needed)
👍 1
w
is anyone using Metaflow with Java/Scala? I'm curious how it's setup -- like is there a separate build system and Metaflow only uses the published jars?
v
you could call Java/Scala apps like any other external binary e.g. if you had an uber-jar that you could call on the CLI