Elasticsearch vs relational databases - Performance Comparison

Time Complexity: Elasticsearch vs relational databases

O(log n)

Understanding Time Complexity

When comparing Elasticsearch and relational databases, it's important to understand how their query times grow as data size increases.

We want to see how the cost of searching or retrieving data changes with bigger datasets.

Scenario Under Consideration

Analyze the time complexity of a simple search query in Elasticsearch and a SQL SELECT query in a relational database.

    // Elasticsearch query example
    GET /products/_search
    {
      "query": {
        "match": { "name": "laptop" }
      }
    }

    -- Equivalent SQL query
    SELECT * FROM products WHERE name LIKE '%laptop%';
    

This code searches for products with "laptop" in their name using Elasticsearch and a relational database.

Identify Repeating Operations

Both queries scan through data to find matches.

• Primary operation: Searching through indexed data or table rows.
• How many times: Depends on the number of documents or rows (n).

How Execution Grows With Input

As the number of items grows, the search time changes differently in each system.

Input Size (n)	Elasticsearch Approx. Operations	Relational DB Approx. Operations
10	Quick, near constant time due to inverted index	May scan many rows, slower
100	Still fast, index helps find matches quickly	Slower, more rows to check
1000	Search time grows slowly, index efficient	Search time grows roughly linearly with rows

Pattern observation: Elasticsearch uses indexes to keep search time low even as data grows, while relational DBs without indexes scan more data and slow down more.

Final Time Complexity

Time Complexity: O(log n) for Elasticsearch with indexes, O(n) for relational DB without indexes

This means Elasticsearch searches grow slowly with data size, while relational DB searches can grow directly with data size if not indexed.

Common Mistake

[X] Wrong: "All database searches take the same time no matter the system."

[OK] Correct: Different systems use different ways to find data. Elasticsearch uses special indexes that make searching faster as data grows, unlike some relational databases without proper indexes.

Interview Connect

Understanding how search time changes with data size helps you explain why certain databases fit different needs. This skill shows you can think about performance, not just code.

Self-Check

"What if the relational database had proper indexes on the searched column? How would the time complexity change?"

Practice

1. Which of the following best describes the main use case for Elasticsearch compared to relational databases?

easy

A. Elasticsearch is optimized for fast full-text search and analytics.

B. Elasticsearch is mainly used for storing structured transactional data.

C. Elasticsearch uses SQL to manage complex relationships between tables.

D. Elasticsearch is designed for batch processing of large datasets only.

Solution

1. Step 1: Understand Elasticsearch's purpose

   Elasticsearch is built for fast full-text search and analytics on large volumes of data.

2. Step 2: Compare with relational databases

   Relational databases focus on structured data and transactional integrity, not optimized for text search.

3. Final Answer:

   Elasticsearch is optimized for fast full-text search and analytics. -> Option A

4. Quick Check:

   Elasticsearch = fast text search [OK]

Hint: Elasticsearch = search and analytics, relational = structured data [OK]

Common Mistakes:

• Confusing Elasticsearch with relational database transactional features
• Thinking Elasticsearch uses SQL for queries
• Assuming Elasticsearch is only for batch processing

2. Which query language is primarily used to interact with relational databases?

easy

A. XPath

B. JSON Query DSL

C. SQL

D. GraphQL

Solution

1. Step 1: Identify relational database query language

   Relational databases use SQL (Structured Query Language) to manage and query data.

2. Step 2: Compare other options

   JSON Query DSL is for Elasticsearch, XPath is for XML, GraphQL is a query language for APIs.

3. Final Answer:

   SQL -> Option C

4. Quick Check:

   Relational DBs use SQL [OK]

Hint: Relational DBs use SQL, Elasticsearch uses JSON DSL [OK]

Common Mistakes:

• Confusing JSON Query DSL with SQL
• Thinking XPath or GraphQL are used for relational databases
• Mixing query languages between technologies

3. Given the following Elasticsearch query snippet, what is the main purpose of this query?

{"query": {"match": {"title": "Elasticsearch"}}}

medium

A. Update documents with the title 'Elasticsearch'.

B. Create a new document with title 'Elasticsearch'.

C. Delete documents where the title is exactly 'Elasticsearch'.

D. Retrieve documents where the title field contains the word 'Elasticsearch'.

Solution

1. Step 1: Analyze the query structure

   The query uses "match" on the "title" field with the term "Elasticsearch" which searches for documents containing that word.

2. Step 2: Understand Elasticsearch query actions

   Queries retrieve data; update, delete, and create require different API calls.

3. Final Answer:

   Retrieve documents where the title field contains the word 'Elasticsearch'. -> Option D

4. Quick Check:

   Match query = search documents [OK]

Hint: "match" queries search text fields in Elasticsearch [OK]

Common Mistakes:

• Thinking the query modifies data instead of searching
• Confusing match with exact match or update operations
• Assuming it deletes or creates documents

4. Identify the error in this SQL query compared to Elasticsearch query style:

SELECT * FROM products WHERE name = 'laptop'

medium

A. SQL uses JSON format, so this query is invalid.

B. Elasticsearch does not use SQL syntax; it uses JSON-based queries.

C. The query is missing a JOIN clause required in Elasticsearch.

D. The query should use double quotes around 'laptop' in SQL.

Solution

1. Step 1: Understand query languages

   SQL uses its own syntax; Elasticsearch uses JSON Query DSL, not SQL syntax.

2. Step 2: Identify the mismatch

   The given query is valid SQL but invalid for Elasticsearch which requires JSON format queries.

3. Final Answer:

   Elasticsearch does not use SQL syntax; it uses JSON-based queries. -> Option B

4. Quick Check:

   Elasticsearch queries = JSON, not SQL [OK]

Hint: Elasticsearch queries are JSON, not SQL [OK]

Common Mistakes:

• Assuming Elasticsearch supports SQL syntax natively
• Thinking JOINs are mandatory in Elasticsearch queries
• Confusing string quoting rules between SQL and JSON

5. You need to build a system that supports complex transactions with multiple table joins and strict data integrity. Which database type should you choose and why?

hard

A. Relational database, because it supports ACID transactions and joins.

B. Relational database, because it is optimized for full-text search.

C. Elasticsearch, because it stores data in JSON and is schema-less.

D. Elasticsearch, because it handles complex joins efficiently.

Solution

1. Step 1: Identify requirements

   The system needs complex transactions, multiple joins, and strict data integrity.

2. Step 2: Match requirements to database features

   Relational databases provide ACID transactions and support joins, ensuring data integrity.

3. Step 3: Compare Elasticsearch capabilities

   Elasticsearch is optimized for search and analytics, not transactional integrity or joins.

4. Final Answer:

   Relational database, because it supports ACID transactions and joins. -> Option A

5. Quick Check:

   Transactions + joins = Relational DB [OK]

Hint: Use relational DB for transactions and joins, Elasticsearch for search [OK]

Common Mistakes:

• Choosing Elasticsearch for transactional systems
• Confusing JSON storage with transactional support
• Assuming Elasticsearch supports complex joins