How to Use Should in Bool Query in Elasticsearch

In Elasticsearch, use the should clause inside a bool query to specify optional conditions that increase the relevance score if matched. Documents matching more should clauses score higher, but if no must clauses exist, at least one should clause must match for the document to be included.

Syntax

The should clause is part of the bool query and contains an array of queries. Each query inside should is optional but boosts the document's score if matched.

Key parts:

• bool: Combines multiple query clauses.
• should: List of optional queries.
• minimum_should_match (optional): Number of should clauses that must match.

json

{
  "query": {
    "bool": {
      "should": [
        { "match": { "field1": "value1" } },
        { "match": { "field2": "value2" } }
      ],
      "minimum_should_match": 1
    }
  }
}

Example

This example searches for documents where either title contains "Elasticsearch" or content contains "search". Documents matching both score higher.

json

{
  "query": {
    "bool": {
      "should": [
        { "match": { "title": "Elasticsearch" } },
        { "match": { "content": "search" } }
      ],
      "minimum_should_match": 1
    }
  }
}

Output

{
  "hits": {
    "total": { "value": 3, "relation": "eq" },
    "hits": [
      { "_score": 1.5, "_source": { "title": "Elasticsearch Basics", "content": "Learn search" } },
      { "_score": 1.2, "_source": { "title": "Advanced Search", "content": "Elasticsearch tips" } },
      { "_score": 0.8, "_source": { "title": "Search Engines", "content": "Introduction" } }
    ]
  }
}

Common Pitfalls

Common mistakes when using should include:

• Not setting minimum_should_match when no must clauses exist, causing unexpected results.
• Expecting should clauses to filter documents strictly; they only influence scoring unless minimum_should_match is set.
• Using should with must clauses without understanding that should clauses become optional.

json

{
  "query": {
    "bool": {
      "should": [
        { "match": { "field": "value1" } },
        { "match": { "field": "value2" } }
      ]
      /* Missing minimum_should_match, no must clause */
    }
  }
}

/* Correct usage with minimum_should_match */
{
  "query": {
    "bool": {
      "should": [
        { "match": { "field": "value1" } },
        { "match": { "field": "value2" } }
      ],
      "minimum_should_match": 1
    }
  }
}

Quick Reference

Summary tips for using should in bool queries:

• Use should to add optional conditions that boost relevance.
• Set minimum_should_match when no must clauses exist to require matches.
• Combine with must for mandatory conditions plus optional boosts.
• Remember should clauses alone do not filter unless minimum_should_match is set.

Key Takeaways

Use should inside bool queries to add optional matching conditions that increase score.

If no must clauses exist, set minimum_should_match to require at least one should match.

should clauses alone do not filter results; they only affect scoring unless minimum_should_match is used.

Combine should with must to mix required and optional conditions.

Avoid forgetting minimum_should_match when using only should clauses to prevent unexpected results.