Search-as-you-type field in Elasticsearch - Mini Project: Build & Apply

Search-as-you-type field

📖 Scenario: You are building a search feature for a website that helps users find products quickly as they type. To make the search fast and helpful, you will create a special field in Elasticsearch called a search-as-you-type field. This field lets Elasticsearch find matches even if the user types only part of a word.

🎯 Goal: Create an Elasticsearch index with a search_as_you_type field called product_name. Then add some sample products. Finally, write a query that finds products as the user types part of the name.

📋 What You'll Learn

Create an index called products with a product_name field of type search_as_you_type

Add three products with exact names: Apple iPhone, Apple Watch, and Samsung Galaxy

Write a search query that uses multi_match with type set to bool_prefix to find products by partial input

Print the names of matching products from the search results

💡 Why This Matters

🌍 Real World

Search-as-you-type fields are used in online stores, libraries, and apps to help users find items quickly by typing only part of the name.

💼 Career

Knowing how to set up and query search-as-you-type fields is useful for backend developers, search engineers, and anyone working with Elasticsearch to build fast and user-friendly search features.

1

Create the Elasticsearch index with a search-as-you-type field

Create an index called products with a mapping that defines a field named product_name of type search_as_you_type. Use the Elasticsearch PUT request with the mapping inside the mappings key.

Elasticsearch

# Create the index 'products' with 'product_name' as search_as_you_type
# Your code here

Need a hint?

Use the PUT method to create the index. Inside mappings, define properties with product_name as search_as_you_type.

2

Add sample products to the index

Add three documents to the products index with the exact product_name values: Apple iPhone, Apple Watch, and Samsung Galaxy. Use the POST method to add each document with an id of your choice.

Elasticsearch

PUT /products
{
  "mappings": {
    "properties": {
      "product_name": {
        "type": "search_as_you_type"
      }
    }
  }
}

# Add document with id 1 and product_name 'Apple iPhone'
# Your code here

Need a hint?

Use POST /products/_doc/<id> to add each product. The product_name field must match exactly.

3

Write a search query using multi_match with bool_prefix

Write a search query using POST /products/_search that uses multi_match with type set to bool_prefix. Search the product_name field for the partial input Apple. This will find products starting with 'Apple'.

Elasticsearch

PUT /products
{
  "mappings": {
    "properties": {
      "product_name": {
        "type": "search_as_you_type"
      }
    }
  }
}

POST /products/_doc/1
{
  "product_name": "Apple iPhone"
}

POST /products/_doc/2
{
  "product_name": "Apple Watch"
}

POST /products/_doc/3
{
  "product_name": "Samsung Galaxy"
}

# Write a search query to find products starting with 'Apple'
# Your code here

Need a hint?

Use multi_match with type set to bool_prefix. Include product_name and its _2gram and _3gram subfields in fields.

4

Print the names of matching products from the search results

Print the product_name values from the search results of the previous query. Show only the names of the products that matched the partial input Apple.

Elasticsearch

PUT /products
{
  "mappings": {
    "properties": {
      "product_name": {
        "type": "search_as_you_type"
      }
    }
  }
}

POST /products/_doc/1
{
  "product_name": "Apple iPhone"
}

POST /products/_doc/2
{
  "product_name": "Apple Watch"
}

POST /products/_doc/3
{
  "product_name": "Samsung Galaxy"
}

POST /products/_search
{
  "query": {
    "multi_match": {
      "query": "Apple",
      "type": "bool_prefix",
      "fields": ["product_name", "product_name._2gram", "product_name._3gram"]
    }
  }
}

# Print the product_name values from the search hits
# Your code here

Need a hint?

Extract the product_name from each hit in the search results and print it. The output should list the matching product names line by line.