Elasticsearch ranks search results by calculating a score for each matching document. This score depends on the indexed data & the particular query. The calculated scores define the order in which the results are returned. The document with the highest score is returned first, the second highest scored is returned second, etc.
❗️ The score defines the order in which results are ranked. The scores themselves have no application level meaning. Given a particular query the calculated scores specify the relation between documents only. Different queries may produce different scores.
Each particular query generates its own range of scores that depends on a number of factors, e.g. sub queries, combinations, boost factors, indexed terms. Similar queries that search in separate fields, e.g. title vs description calculate scores on different sets of indexed terms. Combining these scores into a single list of results will most likely produce unexpected or wrong results, because their score ranges will be different.
🔎 For more information on scoring, see the Elasticsearch Guide on Scoring.
The query context in a query tries to answer question How well does a document match this query clause? For these queries the relevance score is calcualted.
The filter context in a query answers the question Does this document match this query clause? It's a binary decision, either the document matches or not. For these filters no score is calculated or a default value of
1.0is automatically applied.
- Query And Filter Context - Elasticsearch documentation
- How Scoring Works In Elasticsearch