As of now, ACID properties in Apache Hive are only supported for the ORC file format. I have data in Parquet files (lets say they're very large). I want to figure out the pros and cons of leaving the data as is and using a non-ACID table vs converting the Parquet data to ORC (thus having to deal with 2 sets of data and keeping them in sync) and having ACID properties. I wanted to understand which aspects of ACID exactly aren't offered if using a non-ACID table and what limitations it puts on my usage of that data. Any links to further information would also help!
This link tells me about the ACID properties that are guaranteed when using the ORC format. However, I can't find any information as to what would be the impact (ex - Do transactions in a non-ACID table not lock the data? Are certain operations not allowed? etc) if not using ACID.