Storing the bigdata resultset using validation_file_store is very slow for large datasets due to all invalid row indexes.

[vivekgana]

We are running a large dataset with 1 billion rows and 30 expectations, and we need to capture all failed row indexes on a very high-performance setup with a 50-worker cluster. The GX library is serializing write calls to the file and performing very slowly. Has anyone else tried to run all invalid rows to a resultset on a large input file?