ClickHouse 22.3: boosting performance with new features and optimizations

ClickHouse version 25.2 debuts with enhanced features aimed at improving performance, including faster parallel hash joins and the ability to write Bloom filters for Parquet files, which increases query speed by 30-40%. The update also introduces transitive condition inference to enhance filtering efficiency, a backup database engine for instant read-only attachments, and experimental integration with Delta Lake using the Delta Rust Kernel. Additionally, new JSON output formats were introduced to support real-time event streaming over HTTP, offering features like real-time progress bars and dynamic result tables.