CrystallineCore Releases Biscuit Accelerating PostgreSQL LIKE Queries
CrystallineCore launched Biscuit, an open-source PostgreSQL extension on GitHub that implements a custom index access method to accelerate LIKE and ILIKE pattern matching. Using in-memory Roaring Bitmaps and bitmap-based pruning, the project reports dramatic speedups—benchmarks show a wildcard query on 10 million rows falling from over 30 seconds to under 200 milliseconds. The extension targets text-heavy workloads such as e-commerce search and log analytics.
Key Points
- 1Implements custom index access method using in-memory Roaring Bitmaps for native multi-column LIKE/ILIKE pattern matching
- 2Reduces wildcard query latency dramatically — example: 30+ seconds to under 200 milliseconds on 10 million rows
- 3Enables practitioners to accelerate text-heavy searches, lower server costs, and cut ETL job times by ~40%
Scoring Rationale
High, actionable performance gains and official project artifacts, limited by initial compatibility and index-build overhead on large tables.
Sources
Public references used for this report.
Practice with real Telecom & ISP data
90 SQL & Python problems · 15 industry datasets
250 free problems · No credit card
See all Telecom & ISP problems
