Atlas Search

Configure the Atlas Search operator, compound/filter, score, highlight, sequence token, paging/count, and Driver-native extension paths based on version 1.0.0.

Basic Atlas Search

java

$search and $searchMeta must be the first stage of the aggregation pipeline, so they are separated from the general fields(...).end() flow. You can also use .search() without an index.

Search operator

• root builder: text, phrase, autocomplete, equals, exists, in, range, compound

• reusable clause: SearchOperators.text/phrase/autocomplete/equals/in/exists/range

• Driver-native root operator: operator(SearchOperator)

• named Driver wrapper: SearchOperators.driver(name, operator)

java

compound supports must, mustNot, should, filter, minimumShouldMatch, and score, and you can also directly pass a general AtlasSearchOperator.

Text / Phrase / Autocomplete

java

• text: path(s), query(ies), fuzzy, matchCriteria, synonyms, score

• phrase: path(s), query(ies), slop, synonyms, score

• autocomplete: single path, query(ies), tokenOrder, fuzzy, score

• Do not use fuzzy(...) and synonyms(...) simultaneously.

Difference between Search clause and post-search fields

java

search().fields(...) is a general $match after $search, not an internal filter of $search. To limit candidates at the search index stage, use compound.filter(...).

text

Score and Sorting

java

scoreDesc()/scoreAsc() and general sort determine priority in the order they are called. For stable pagination, it is advisable to include stable unique field sorting along with the score.

Highlight / score details / sequence token

java

The default aliases are highlights, scoreDetails, searchSequenceToken. addFieldsScoreDetails() also activates score details in the Search option.

Paging / sequence token / count

java

count().execute()

The final pipeline count reflecting post-search fields(...) and score threshold.

count().executeSearchMeta()

$searchMeta metadata count. Supports SearchCountType.TOTAL/LOWER_BOUND and does not execute post-search $match.

existsQuery().execute()

Checks if there is more than one result in the Search pipeline.

Driver-native extension

java

stage(Bson) is placed immediately after $search and is not applied to the $searchMeta metadata-count pipeline of executeSearchMeta(). If new Driver features are added before the DSL convenience API, prioritize using operator, driverOptions, stage, SortSpec.driver, and customizeAggregation.