Why Constant score query in Elasticsearch? - Purpose & Use Cases

The Big Idea

Discover how to make your searches lightning-fast by skipping complicated scoring!

The Scenario

Imagine you want to find all documents about "coffee" in a huge database, but you don't care about how relevant each document is, just that it matches. Doing this manually means writing complex queries and then adjusting scores yourself.

The Problem

Manually adjusting scores or relevance is slow and tricky. You might spend hours tweaking weights and still get inconsistent results. It's easy to make mistakes and hard to maintain.

The Solution

The constant score query lets you wrap any filter and assign the same score to all matching documents. This means you get fast, simple results without worrying about relevance calculations.

Before vs After

✗ Before

GET /_search
{
  "query": {
    "bool": {
      "filter": { "term": { "type": "coffee" } },
      "must": { "match": { "description": "coffee" } }
    }
  }
}

✓ After

GET /_search
{
  "query": {
    "constant_score": {
      "filter": { "term": { "type": "coffee" } },
      "boost": 1.0
    }
  }
}

What It Enables

You can quickly filter documents with a fixed score, making searches faster and simpler when relevance doesn't matter.

Real Life Example

For example, an online store wants to show all products in the "coffee" category equally, without ranking by popularity or reviews.

Key Takeaways

Manual scoring is complex and error-prone.

Constant score query assigns the same score to all matches.

This simplifies filtering when relevance is not needed.

Practice

1. What is the main purpose of a constant_score query in Elasticsearch?

easy

A. To filter documents and assign the same score to all matched documents

B. To sort documents by their relevance score

C. To perform a full-text search with varying scores

D. To aggregate documents by a field value

Solution

1. Step 1: Understand the role of constant_score query

   The constant_score query filters documents and assigns the same score to all matches, ignoring relevance scoring.

2. Step 2: Compare with other query types

   Unlike full-text queries that score by relevance, constant_score keeps scores equal for filtered documents.

3. Final Answer:

   To filter documents and assign the same score to all matched documents -> Option A

4. Quick Check:

   constant_score = equal scoring for filtered docs [OK]

Hint: Constant score means all matched docs get the same score [OK]

Common Mistakes:

• Thinking it sorts by relevance
• Confusing with full-text scoring
• Assuming it aggregates data

2. Which of the following is the correct syntax to create a constant score query with a filter on field status equal to active?

easy

A. {"constant_score": {"filter": {"match": {"status": "active"}}}}

B. {"constant_score": {"query": {"term": {"status": "active"}}}}

C. {"constant_score": {"term": {"status": "active"}}}

D. {"constant_score": {"filter": {"term": {"status": "active"}}}}

Solution

1. Step 1: Identify correct key for filtering

   The constant_score query uses the key "filter" to specify the filter clause, not "query" or direct term.

2. Step 2: Check filter syntax

   The filter should be a term query inside the filter object, correctly written as {"term": {"status": "active"}}.

3. Final Answer:

   {"constant_score": {"filter": {"term": {"status": "active"}}}} -> Option D

4. Quick Check:

   Use "filter" with term inside constant_score [OK]

Hint: Use "filter" key with term query inside constant_score [OK]

Common Mistakes:

• Using "query" instead of "filter"
• Placing term query directly under constant_score
• Using match instead of term for exact filter

3. Given this Elasticsearch query snippet, what will be the score of each matched document?

{
  "constant_score": {
    "filter": { "term": { "category": "books" } },
    "boost": 3
  }
}

medium

A. The score will vary based on relevance

B. The score will be 3 for all matched documents

C. The score will be 1 for all matched documents

D. The score will be 0 because boost is ignored

Solution

1. Step 1: Understand constant_score with boost

   The constant_score query assigns the same score to all matched documents, and the boost multiplies that score.

2. Step 2: Calculate the score

   By default, constant_score gives score 1, multiplied by boost 3, resulting in score 3 for all matches.

3. Final Answer:

   The score will be 3 for all matched documents -> Option B

4. Quick Check:

   constant_score x boost = 3 [OK]

Hint: Boost multiplies constant score, so score = boost value [OK]

Common Mistakes:

• Assuming scores vary by relevance
• Ignoring the boost value
• Thinking boost is ignored in constant_score

4. Identify the error in this constant score query:

{
  "constant_score": {
    "filter": {
      "terms": {
        "tags": ["tech", "science"]
      }
    },
    "boost": "2"
  }
}

medium

A. The boost value should be a number, not a string

B. The filter key should be replaced with query

C. The terms query cannot be used inside constant_score

D. The tags field should be inside a match query

Solution

1. Step 1: Check boost data type

   The boost parameter must be a numeric value, not a string. Here, "2" is a string, which is invalid.

2. Step 2: Validate filter and terms usage

   Using "filter" with a "terms" query is valid syntax, so no error there.

3. Final Answer:

   The boost value should be a number, not a string -> Option A

4. Quick Check:

   boost must be numeric, not string [OK]

Hint: Boost must be a number, not quoted string [OK]

Common Mistakes:

• Using string for boost instead of number
• Replacing filter with query incorrectly
• Misunderstanding terms query usage

5. You want to find all documents where status is published and boost their score by 5, but keep all matched documents scored equally. Which query correctly achieves this?

hard

A. { "match": { "status": "published", "boost": 5 } }

B. { "bool": { "must": { "term": { "status": "published" } }, "boost": 5 } }

C. { "constant_score": { "filter": { "term": { "status": "published" } }, "boost": 5 } }

D. { "constant_score": { "query": { "term": { "status": "published" } }, "boost": 5 } }

Solution

1. Step 1: Use constant_score with filter and boost

   To keep scores equal and apply a boost, use constant_score with a filter clause and boost parameter.

2. Step 2: Validate correct syntax

   { "constant_score": { "filter": { "term": { "status": "published" } }, "boost": 5 } } uses "filter" with a term query and numeric boost, which is the correct syntax.

3. Step 3: Check other options

   { "bool": { "must": { "term": { "status": "published" } }, "boost": 5 } } uses bool query which scores by relevance, not constant. { "match": { "status": "published", "boost": 5 } } is a match query with boost but scores vary. { "constant_score": { "query": { "term": { "status": "published" } }, "boost": 5 } } incorrectly uses "query" instead of "filter" inside constant_score.

4. Final Answer:

   { "constant_score": { "filter": { "term": { "status": "published" } }, "boost": 5 } } -> Option C

5. Quick Check:

   constant_score + filter + boost = equal boosted scores [OK]

Hint: Use constant_score with filter and numeric boost for equal boosted scores [OK]

Common Mistakes:

• Using query instead of filter inside constant_score
• Using bool or match queries that score variably
• Setting boost as string instead of number