A previous article presented a performance comparison between PostgreSQL Full-Text Search (FTS) and plain pattern matching, especially in cases where a given piece of text is searched across multiple attributes of multiple entities. In brief, the conclusion was that FTS is significantly faster, while pattern matching offers greater flexibility, a wider range of applicability, and higher precision.

The present article aims to continue this analysis, which is worthwhile when implementing global search solutions, in order to increase objectivity regarding the applicability of these methods and to help programmers make informed choices more easily.