System and method for efficient connection management in a massively parallel or distributed database environment

A system and method is described for database split generation in a massively parallel or distributed database environment including a plurality of databases and a data warehouse layer providing data summarization and querying functionality. A database table accessor obtains split generation information from the database.