mph-table

Uses minimal perfect hashing to reduce memory footprint, as highlighted in the README's focus on efficient space utilization for large datasets.

Optimized for quick key lookups and iteration, ideal for read-heavy workloads, demonstrated by the fast reads feature in the README.

Designed for offline table construction via batch processes, separating writes and reads to optimize performance, as per the philosophy section.

Supports configurable serializers for keys and values, including complex types like lists and dictionaries, shown in the command-line examples with nested serializers.

Provides both a Java library for programmatic access and command-line interfaces for easy table creation and inspection without coding, as mentioned in the usage section.