Apache Druidvspgvector

Pros

• The database is a real-time analytics database designed for high-performance slice-and-dice queries on massive data sets, often delivering sub-second response times even on trillions of records .
• It features a scalable, distributed architecture with separate services for ingestion, storage, and querying that can be independently scaled to handle high ingest rates of millions of records per second .
• The columnar storage format and compressed bitmap indexes enable fast filtering and aggregation by only loading the columns needed for a query, which greatly improves query performance .

Cons

• While Druid excels at streaming inserts, it does not support low-latency updates of existing records using a primary key; updates must be performed using background batch jobs .
• The database is optimized for queries that hit one large distributed table; it is not ideal for "big joins" between two large fact tables, which can be slow and inefficient .
• The platform is most effective for time-series or event-based data with a time component; its architecture and optimizations may not be a good fit for general-purpose transactional workloads or for use cases where query latency is not a critical concern .

Pros

Cons