Solving Postgres' Count Problem
Summary
ParadeDB's blog post 'Making COUNTs and the Rest of Search Really Fast (Part II)' explains optimizations to speed up COUNT queries in Postgres. It details segment partitioning by ctids, inverting the DocId map, and using bitmaps for dense postings, with benchmark comparisons against TIN on Wikipedia and Hacker News datasets. The piece discusses architectural differences and benchmarking methodology, and invites readers to test ParadeDB on their own data.