Index Searching with Regular Expressions
summary
Describes building an index to speed up regex search over large text collections by combining finite automata with inverted-index techniques. Explains using automaton-based filtering (NFA/DFA), extracting fixed substrings or k-grams from patterns to query an inverted index, and then verifying matches to reduce full-regex evaluations.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 1:30Regex and finite automata
- 4:17Regex search problem
- 6:45Inverted indexes and n-grams
- 9:40Regex to logical expressions
- 12:10Yandex Fast Text Express
- 14:24Google Search approach
- 17:54Speaker's method: automaton and trigrams
- 22:12Method comparison
- 24:24Performance tests
- 25:50Problems and limitations
- 27:24SQL integration
- 30:00Optimization and future work
- 32:16Q&A
related talks
45mSQL Indexing Beyond Simple Columns: Expressions, Text Fragments, Attributes J...
51mSQL indexing: expressions, text fragments, JSON attributes and Top-N queries
37mAlgorithms & data-structures that power Lucene & ElasticSearch
49mBreakout Session
26mFrom __icontains to search
31mUnspoken overhead of the Inverted Index
24mSearch From the Ground Up
26mOn The Look-Out For Your Data
