Bool query in depth in Elasticsearch

Introduction

A bool query helps you combine many search rules together. It lets you say what must happen, what should happen, and what must not happen in your search.

You want to find documents that match several conditions at once.

You want to require some conditions but make others optional.

You want to exclude documents that have certain words or values.

You want to mix different types of queries like term, range, or match.

You want to build complex search filters for your app or website.

Syntax

Elasticsearch

{
  "bool": {
    "must": [ /* queries that must match */ ],
    "filter": [ /* queries that filter results but don't affect score */ ],
    "should": [ /* queries that should match to boost score */ ],
    "must_not": [ /* queries that must NOT match */ ]
  }
}

must means all these queries have to match.

filter is like must but faster and does not change score.

Examples

Find documents where the title has the word "apple".

Elasticsearch

{
  "bool": {
    "must": [
      { "match": { "title": "apple" } }
    ]
  }
}

Find documents with "apple" in the title but exclude those with status "sold".

Elasticsearch

{
  "bool": {
    "must": [
      { "match": { "title": "apple" } }
    ],
    "must_not": [
      { "term": { "status": "sold" } }
    ]
  }
}

Find documents that mention either "fresh" or "organic" in the description.

Elasticsearch

{
  "bool": {
    "should": [
      { "match": { "description": "fresh" } },
      { "match": { "description": "organic" } }
    ],
    "minimum_should_match": 1
  }
}

Filter documents where price is between 10 and 50, without affecting score.

Elasticsearch

{
  "bool": {
    "filter": [
      { "range": { "price": { "gte": 10, "lte": 50 } } }
    ]
  }
}

Sample Program

This query finds documents that have "coffee" in the title and "fresh" in the description. It only includes items priced between 5 and 20. It excludes items with status "sold". If the brand is "acme", the document scores higher but it is not required.

Elasticsearch

{
  "query": {
    "bool": {
      "must": [
        { "match": { "title": "coffee" } },
        { "match": { "description": "fresh" } }
      ],
      "filter": [
        { "range": { "price": { "gte": 5, "lte": 20 } } }
      ],
      "must_not": [
        { "term": { "status": "sold" } }
      ],
      "should": [
        { "match": { "brand": "acme" } }
      ],
      "minimum_should_match": 0
    }
  }
}

Important Notes

You can use minimum_should_match to say how many should queries must match.

filter clauses are cached and faster for repeated queries.

Use must_not to exclude unwanted results.

Summary

A bool query combines many queries with rules: must, should, filter, and must_not.

Use must for required matches, filter for fast filtering, should for optional boosts, and must_not to exclude.

Bool queries help build powerful and flexible searches in Elasticsearch.

Practice

1. Which clause in an Elasticsearch bool query is used to exclude documents from the search results?

easy

A. must

B. should

C. must_not

D. filter

Solution

1. Step 1: Understand bool query clauses

   The bool query has four main clauses: must (required matches), filter (fast filtering), should (optional matches), and must_not (exclusions).

2. Step 2: Identify exclusion clause

   The must_not clause excludes documents that match its conditions from the results.

3. Final Answer:

   must_not -> Option C

4. Quick Check:

   Exclusion = must_not [OK]

Hint: Exclude with must_not clause in bool query [OK]

Common Mistakes:

• Confusing must_not with should
• Using filter to exclude documents
• Thinking must excludes documents

2. Which of the following is the correct syntax to add a filter clause inside a bool query in Elasticsearch?

easy

A. "bool": { "must": [ { "term": { "status": "active" } } ] }

B. "bool": { "must_not": { "term": { "status": "active" } } }

C. "bool": { "should": { "term": { "status": "active" } } }

D. "bool": { "filter": [ { "term": { "status": "active" } } ] }

Solution

1. Step 1: Recall bool query syntax for filter

   The filter clause expects an array of queries inside the bool query, e.g., "filter": [ { ... } ].

2. Step 2: Match correct syntax

   "bool": { "filter": [ { "term": { "status": "active" } } ] } correctly uses "filter": [ { "term": { "status": "active" } } ] which is valid syntax.

3. Final Answer:

   "bool": { "filter": [ { "term": { "status": "active" } } ] } -> Option D

4. Quick Check:

   Filter clause uses array syntax [OK]

Hint: Filter clause always uses an array of queries [OK]

Common Mistakes:

• Using object instead of array for filter
• Confusing must_not or should syntax
• Missing brackets around filter queries

3. Given this bool query snippet, what documents will it match?

{ "bool": { "must": [ { "match": { "title": "python" } } ], "should": [ { "term": { "tags": "programming" } } ], "must_not": [ { "term": { "status": "draft" } } ] } }

medium

A. Documents with title containing 'python', optionally tagged 'programming', excluding drafts

B. Documents tagged 'programming' only, ignoring title and status

C. Documents with status 'draft' and title 'python'

D. Documents with title 'python' or status 'draft'

Solution

1. Step 1: Analyze must clause

   The must clause requires documents to have 'python' in the title.

2. Step 2: Analyze should and must_not clauses

   The should clause boosts documents tagged 'programming' but is optional. The must_not clause excludes documents with status 'draft'.

3. Final Answer:

   Documents with title containing 'python', optionally tagged 'programming', excluding drafts -> Option A

4. Quick Check:

   Must + should + must_not = Documents with title containing 'python', optionally tagged 'programming', excluding drafts [OK]

Hint: Must is required, should is optional, must_not excludes [OK]

Common Mistakes:

• Thinking should is required
• Including documents with status 'draft'
• Ignoring must clause

4. Identify the error in this bool query snippet:

{ "bool": { "must": { "term": { "user": "kimchy" } }, "filter": { "term": { "status": "active" } } } }

medium

A. Missing "should" clause

B. "must" and "filter" should be arrays, not single objects

C. Using "term" inside bool is invalid

D. The query is correct as is

Solution

1. Step 1: Check bool clause types

   In bool queries, must and filter expect arrays of queries, not single objects.

2. Step 2: Identify syntax error

   Here, both must and filter are single objects, which is invalid syntax.

3. Final Answer:

   "must" and "filter" should be arrays, not single objects -> Option B

4. Quick Check:

   Bool clauses require arrays [OK]

Hint: Bool clauses must and filter need arrays [OK]

Common Mistakes:

• Using single object instead of array
• Forgetting to wrap queries in []
• Assuming should is mandatory

5. You want to find documents where the category is 'books' and either the author is 'Alice' or the year is after 2015, but exclude any documents marked as archived. Which bool query structure correctly implements this?

hard

A. { "bool": { "must": [ { "term": { "category": "books" } } ], "should": [ { "term": { "author": "Alice" } }, { "range": { "year": { "gt": 2015 } } } ], "must_not": [ { "term": { "status": "archived" } } ] } }

B. { "bool": { "filter": [ { "term": { "category": "books" } }, { "term": { "author": "Alice" } } ], "must_not": [ { "term": { "status": "archived" } } ], "should": [ { "range": { "year": { "gt": 2015 } } } ] } }

C. { "bool": { "must": [ { "term": { "category": "books" } }, { "term": { "author": "Alice" } } ], "must_not": [ { "term": { "status": "archived" } } ], "should": [ { "range": { "year": { "gt": 2015 } } } ] } }

D. { "bool": { "must": [ { "term": { "category": "books" } } ], "filter": [ { "term": { "author": "Alice" } }, { "range": { "year": { "gt": 2015 } } } ], "must_not": [ { "term": { "status": "archived" } } ] } }

Solution

1. Step 1: Identify required and optional conditions

   The category must be 'books' (required), and either author is 'Alice' or year > 2015 (optional, so use should).

2. Step 2: Exclude archived documents

   Use must_not to exclude documents with status 'archived'.

3. Step 3: Match query structure

   { "bool": { "must": [ { "term": { "category": "books" } } ], "should": [ { "term": { "author": "Alice" } }, { "range": { "year": { "gt": 2015 } } } ], "must_not": [ { "term": { "status": "archived" } } ] } } correctly uses must for category, should for author or year, and must_not for exclusion.

4. Final Answer:

   Option A's bool query structure -> Option A

5. Quick Check:

   Must + should + must_not for conditions [OK]

Hint: Use must for required, should for alternatives, must_not to exclude [OK]

Common Mistakes:

• Putting author and year in must (both required)
• Using filter instead of should for alternatives
• Missing must_not for exclusion