Write a function to identify duplicate records in a large dataset without using built-in deduplication libraries.
