job.yaml

job:
  name: transform-user-events
  actions:
    - name: load users
      actor:
        type: file
        properties:
          format: csv
          options:
            header: false
            delimiter: ","
            quote: \"
            timestampFormat: "yyyy/MM/dd HH:mm:ss"
          ddlSchemaString: "user_id long, birth_year int, gender string, location string"
          fileUri: "${events.users_input}"
      output-view:
        name: users
        global: true
    - name: load events
      actor:
        type: flat
        properties:
          fileUri: "${events.events_input}"
      output-view:
        name: events_raw
        global: false
    - name: transform-events
      actor:
        type: sql
        properties:
          sqlString: >
            select
              substr(row_value, 1, 12) as event_id,
              substr(row_value, 13, 16) as event_time,
              substr(row_value, 29, 12) as event_host,
              substr(row_value, 41, 64) as event_location
            from events_raw
            where row_no not in (1, 2) and substr(row_value, 6, 5) != 'TFYKR'
      input-views:
        - events_raw
      output-view:
        name: events
        global: true