From Sparkl SQL:
Spark SQL does not support concurrency control because it is designed for read-only OLAP. It provides an interface to save DataFrames to data sources like parquet, or persist tables. It has different save modes for the way handling existing data if present. But these save modes do not utilize any locking and are not atomic.
https://github.com/sameeragarwal/blinkdb
University of California-Berkeley, Massachusetts Institute of Technology
2012
2014