The goal of a Cassandra Batch statement is to group statements on the one partition together in a single atomic operation (all pass or fail together).
This is quite similar to how an SQL Transaction works.
To me it seems the main difference is that with an SQL Transaction, you get an ACID consistency guarantee at the end of it, which you don't necessarily get with a Cassandra Batch statement.
My question is: What are the similarities and differences between a BATCH statement in Cassandra and a Transaction in SQL?