Elasticsearch Course

19 lessons · free to read

Model, index and query text properly - and know why your term query returns nothing.

Take this course in the learning hub

Lessons

• Why a LIKE query is not searchA relational scan and a search engine answer two genuinely different questions.
• The inverted indexThe one data structure that makes full-text search fast, and what it does not do.
• Documents, indices and near real timeWhy a document you just indexed is not in the next search, and when it will be.
• Shards, replicas and cluster healthHow an index is split across nodes, and what yellow and red are actually telling you.
• Analyzers, tokenizers and token filtersThe chain that turns a string into terms, and why it must be the same on both sides.
• Mapping and the dynamic mapping trapDynamic mapping is convenient exactly until the first document guesses wrong.
• text versus keywordThe single most common Elasticsearch mistake, and the multi-field that avoids it.
• match, term and match_phraseThree queries that look interchangeable and behave completely differently.
• bool queries and filter contextCombine clauses correctly, and stop paying for scores you are going to ignore.
• Relevance and BM25Why these ten documents came back in this order, and how to make the engine tell you.
• AggregationsTurn the matching documents into counts, averages and buckets in the same request.
• Deep paginationWhy page 500 fails, and the two supported ways of walking a large result set.
• Rollover and index lifecycle managementTime-based data belongs in a rolling series of indices, managed by a policy.
• Elasticsearch from Spring BootWire the client, map the entity, and know when to drop to a native query.
• Shard sizing and overshardingThe decision you cannot change later, and the one that destroys clusters.
• Filter context, caching and the slow query pathsThe same result at very different cost, depending on how you ask.
• Mapping explosion and field countDynamic mapping plus unpredictable keys is how an index becomes unusable.
• Aggregations, cardinality and circuit breakersThe aggregations that are cheap, and the ones that take a node down.
• How an Elasticsearch cluster actually failsYellow, red, split brain and the disk watermark that makes an index read-only.