Text vs keyword field types in Elasticsearch - Practice Questions

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of this Elasticsearch query?

Given the following mapping and query, what will be the result count?

{
  "mappings": {
    "properties": {
      "name": { "type": "text" },
      "category": { "type": "keyword" }
    }
  }
}

POST /products/_search
{
  "query": {
    "term": { "category": "electronics" }
  }
}

A Returns no documents because 'term' query does not work on keyword fields

B Returns all documents where the category exactly matches 'electronics'

C Returns documents where the name contains the word 'electronics'

D Returns documents where the category contains the word 'electronics' anywhere

💡 Hint

Remember that keyword fields are not analyzed and match exact values.

The term query matches exact values on keyword fields. Since category is a keyword field, it matches documents where the category is exactly 'electronics'.

🧠 Conceptual

intermediate

Why use 'text' type instead of 'keyword' for a field?

Which reason best explains why you would choose text type over keyword type for a field in Elasticsearch?

A To enable full-text search with tokenization and analysis

B To store exact values for filtering and sorting

C To reduce index size by storing unanalyzed data

D To improve performance of term queries

💡 Hint

Think about when you want to search inside the text rather than match exact values.

The text type is used for full-text search. It breaks the text into tokens and applies analysis, allowing searches for words inside the text. The keyword type stores exact values without analysis.

🔧 Debug

advanced

Why does this term query return zero results?

Given this mapping and query, why does the term query return zero results?

{
  "mappings": {
    "properties": {
      "description": { "type": "text" }
    }
  }
}

POST /items/_search
{
  "query": {
    "term": { "description": "fast car" }
  }
}

A Because the term query syntax is incorrect and should use 'match' instead

B Because the index is empty and has no documents

C Because the field 'description' is missing from the mapping

D Because 'description' is a text field and term queries require exact matches, so 'fast car' as a whole does not match

💡 Hint

Think about how term queries work on text fields.

Term queries look for exact matches. Since 'description' is a text field, it is analyzed and broken into tokens like 'fast' and 'car'. The term query with 'fast car' as a whole string does not match any token exactly, so it returns zero results.

📝 Syntax

advanced

Which mapping snippet correctly defines a field for exact filtering and full-text search?

Choose the correct mapping snippet that allows a field to be searched as full text and also filtered exactly.

A { "title": { "type": "text", "fields": { "raw": { "type": "keyword" } } } }

B { "title": { "type": "keyword", "fields": { "raw": { "type": "text" } } } }

C { "title": { "type": "text" } }

D { "title": { "type": "keyword" } }

💡 Hint

Think about multi-fields and how to support both search and filtering.

Option A defines a text field for full-text search and a subfield raw of type keyword for exact filtering and sorting. This is the recommended pattern.

🚀 Application

expert

How many unique values can be efficiently filtered using a keyword field?

In Elasticsearch, what is the practical limit on the number of unique values in a keyword field before performance degrades significantly?

A Keyword fields should only have a single unique value for best performance

B There is no limit; keyword fields handle millions of unique values efficiently

C Up to a few thousand unique values is efficient; beyond that, performance may degrade

D Keyword fields are not used for filtering unique values

💡 Hint

Consider how Elasticsearch stores keyword terms and the impact on memory.

Keyword fields store terms in memory for filtering and aggregations. Having too many unique values (cardinality) can cause memory and performance issues. Usually, a few thousand unique values is manageable.