Skip to main content

Search and Count

Note

REST endpoints: _search and _count.

POST /user_info/_search {"query":{"term":{"uid":"1001"}},"size":10}

Map uid as keyword. Results contain _ID and _DOC, with document fields expanded by default. Read a specific field by column name; see Reading Results.

Count​

POST /user_info/_count {"query":{"term":{"uid":"1001"}}}

Use _count for the total, not the number of rows in a search page. See Pagination for dbVisitor Page usage.

Aggregations​

Single-value aggregations return one row, using aggregation names as column names:

POST /user_info/_search {"aggs":{"max_age":{"max":{"field":"age"}}}}

Use composite to group by fields. This example returns age and cnt columns:

POST /user_info/_search {
  "aggs": {
    "rows": {
      "composite": {"sources": [{"age": {"terms": {"field": "age", "missing_bucket": true}}}]},
      "aggs": {"cnt": {"filter": {"match_all": {}}}}
    }
  }
}

Remove the nested aggs to read distinct field combinations; list additional fields in sources. The driver fetches subsequent pages through after_key as needed. JDBC fetchSize controls the number of buckets per page.

For an exact distinct count, add "meta":{"dbvisitor":{"mode":"count","column":"distinct_count"}} inside rows. This traverses every distinct bucket and excludes combinations containing null; it is not a low-cost approximate count.

Search results do not expose the complete REST response, such as _score or highlight.

Scripted Columns​

Names declared in script_fields can be read as result columns. This example returns an age_twice column. A projection containing only scripted columns does not require index-mapping access:

POST /user_info/_search {
  "_source": false,
  "script_fields": {
    "age_twice": {"script": {"source": "doc['age'].value * params.factor", "params": {"factor": 2}}}
  }
}

Bind dynamic values in params, rather than concatenating them into source. Single values return scalars, multiple values return a List, and empty values return null.