Data Cooker ETL is a batch/interactive ETL processor that provides simple yet powerful SQL-like language to perform transformations of columnar and/or structured data, which doesn't require strict schema nor data catalog. Built on Apache Spark core with minimal dependencies.
Historically it was used in a GIS project, so it has geospatial extensions to its SQL, and has been extensively tested and proven reliable by years and petabytes of transformed data in production.
Your contributions are welcome. Please see the source code in the original GitHub repo at https://github.com/PastorGL/datacooker-etl