In my previous post, I demonstrated that Datalevin performs complex queries faster than PostgreSQL. A common reaction is, "Oh, have you tested writing speed? I imagine that when there are indices for everything (as in the case of Datalevin), writing speed could be significantly slower." This post aims to address …

This is a post about Datalevin, an open-source database system that I have been building since 2020. Although JOB here is really an acronym for Join Order Benchmark, a benchmark for complex database queries, I intend to build up Datalevin to compete for the same kind of jobs usually held …

Here is the story. I am adding full-text search capability to Datalevin, a Datalog database that we open sourced last year. For this task, I have decided to write a search engine from scratch instead of using an existing search library. Here are some rationales for this decision. Today I …