job.yml

$schema: https://azuremlschemas.azureedge.net/latest/commandJob.schema.json
code: src
command: >-
  python prep-nyctaxi.py
  --nyc_taxi_dataset ${{inputs.nyc_taxi_dataset}}
inputs:
  nyc_taxi_dataset:
    type: uri_folder
    path: wasbs://nyctlc@azureopendatastorage.blob.core.windows.net/yellow
    mode: ro_mount
environment: 
  image: mcr.microsoft.com/mmlspark/release:1.0.0-rc3
  conda_file: conda.yml
compute: azureml:cpu-cluster-lg
resources:
  instance_count: 1
display_name: spark-nyctaxi-example
experiment_name: spark-nyctaxi-example
description: This sample shows how to run a single node Spark job on Azure ML. The 47GB NYC Taxi dataset is read in parquet format by a 1 node Spark cluster, processed and then written as job output in parquet format.