This project is our collection of Scala modules.
- IntelliJ IDEA
- IntelliJ Scalafmt Plugin
- Scala 2.11
- Spark 2.4.0+
Spark is expected to be a provided dependency, so you should have a working Spark install somewhere, and $SPARK_HOME should be set in your environment.
You should use IntelliJ IDEA (CE is fine). We use the scalafmt IntelliJ IDEA plugin, configured to update on file save, and scalastyle
Some editor config to put in place: Case Class Definition Style
We follow the Twitter Effective Scala style guide.
Saving this here for future reference: Spark + S3
Use homebrew:
brew install [email protected]
brew install sbt
- Spark + S3: http://deploymentzone.com/2015/12/20/s3a-on-spark-on-aws-ec2/
- Spark reading from S3: https://tech.kinja.com/how-not-to-pull-from-s3-using-apache-spark-1704509219
- Spark + GCS: https://stackoverflow.com/a/56400126/11295366
Getting AWS S3 to play nice with Spark is complicated, because it involves a dependency on both aws-java-sdk and hadoop-aws, and these two libraries need to be compatible versions (and compatible with Spark) or else everything explodes:
We currently use AWS 1.7.4 and hadoop-aws 2.7.1 as these are known to be compatible and work with Spark 2.4.0+