Why documents are the unit of data in Elasticsearch - Performance Analysis

Time Complexity: Why documents are the unit of data

O(n)

Understanding Time Complexity

When working with Elasticsearch, documents are the main pieces of data we store and search.

We want to understand how the time it takes to handle data grows as we add more documents.

Scenario Under Consideration

Analyze the time complexity of indexing multiple documents.

POST /my_index/_bulk
{ "index": { "_id": "1" } }
{ "name": "Alice", "age": 30 }
{ "index": { "_id": "2" } }
{ "name": "Bob", "age": 25 }
{ "index": { "_id": "3" } }
{ "name": "Carol", "age": 27 }
    

This code adds several documents to an index in bulk, each document representing one unit of data.

Identify Repeating Operations

Look at what repeats when adding documents.

• Primary operation: Indexing each document one by one.
• How many times: Once per document added.

How Execution Grows With Input

As you add more documents, the work grows with the number of documents.

Input Size (n)	Approx. Operations
10	10 indexing operations
100	100 indexing operations
1000	1000 indexing operations

Pattern observation: The time grows directly with the number of documents added.

Final Time Complexity

Time Complexity: O(n)

This means the time to index data grows in a straight line as you add more documents.

Common Mistake

[X] Wrong: "Adding more documents takes the same time no matter how many there are."

[OK] Correct: Each document needs its own processing, so more documents mean more work and more time.

Interview Connect

Understanding how data size affects processing time helps you explain how Elasticsearch handles scaling in real projects.

Self-Check

"What if we indexed documents in parallel instead of one by one? How would the time complexity change?"

Practice

1. Why are documents considered the unit of data in Elasticsearch?

easy

A. Because documents cannot be searched individually

B. Because documents are only used for storing metadata

C. Because documents are the same as database tables

D. Because each document holds all related information together in one place

Solution

1. Step 1: Understand the role of documents in Elasticsearch

   Documents store all related data fields together, making them self-contained units.

2. Step 2: Compare documents to other data units

   Unlike tables or metadata, documents hold actual data and can be searched independently.

3. Final Answer:

   Because each document holds all related information together in one place -> Option D

4. Quick Check:

   Documents = self-contained data units [OK]

Hint: Remember: documents bundle related data together [OK]

Common Mistakes:

• Thinking documents are just metadata
• Confusing documents with tables
• Assuming documents can't be searched alone

2. Which of the following is the correct way to identify a document in Elasticsearch?

easy

A. Using the document's file path on disk

B. Using the document's creation timestamp only

C. Using a unique document ID within an index

D. Using the document's size in bytes

Solution

1. Step 1: Identify how documents are stored

   Documents are stored inside indexes and each has a unique ID for identification.

2. Step 2: Eliminate incorrect options

   File paths, timestamps, or size are not used as unique identifiers in Elasticsearch.

3. Final Answer:

   Using a unique document ID within an index -> Option C

4. Quick Check:

   Unique ID = document identifier [OK]

Hint: Unique ID inside index identifies documents [OK]

Common Mistakes:

• Confusing file path with document ID
• Using timestamps as unique keys
• Assuming size identifies documents

3. Given an Elasticsearch index with documents representing books, which statement best explains why documents are efficient for searching?

medium

A. Because documents require manual indexing before searching

B. Because each document contains all book details, searches can target whole books quickly

C. Because documents only store keywords, not full data

D. Because documents are stored as separate files on disk, making file search faster

Solution

1. Step 1: Understand document structure for books

   Each document holds all details about a book, like title, author, and summary.

2. Step 2: Analyze search efficiency

   Having all data in one document allows Elasticsearch to quickly find relevant books without joining data.

3. Final Answer:

   Because each document contains all book details, searches can target whole books quickly -> Option B

4. Quick Check:

   Complete data per document = fast search [OK]

Hint: Whole data in one document speeds search [OK]

Common Mistakes:

• Thinking documents are separate files
• Believing documents store only keywords
• Assuming manual indexing is needed every time

4. You have an Elasticsearch document missing its unique ID. What problem will this cause?

medium

A. The document cannot be uniquely identified or updated

B. The document will be stored twice automatically

C. The document will be ignored during searches

D. The document will be converted into multiple smaller documents

Solution

1. Step 1: Understand the role of unique IDs

   Unique IDs allow Elasticsearch to find, update, or delete specific documents.

2. Step 2: Identify the impact of missing ID

   Without an ID, Elasticsearch cannot uniquely identify the document, causing update or delete failures.

3. Final Answer:

   The document cannot be uniquely identified or updated -> Option A

4. Quick Check:

   Missing ID = no unique identification [OK]

Hint: Missing ID means no unique document tracking [OK]

Common Mistakes:

• Assuming document duplicates automatically
• Thinking document is ignored in search
• Believing document splits into parts

5. You want to design an Elasticsearch index for storing user profiles. Why should each profile be a separate document rather than splitting data across multiple documents?

hard

A. Because storing all profile data in one document makes retrieval and updates simpler and faster

B. Because multiple documents per profile improve search speed by splitting data

C. Because Elasticsearch cannot store more than one document per user

D. Because documents must always be very small in size

Solution

1. Step 1: Consider data grouping in Elasticsearch

   Each document should hold all related data for easy access and management.

2. Step 2: Evaluate impact on retrieval and updates

   One document per profile means fewer queries and simpler updates, improving performance.

3. Step 3: Reject incorrect assumptions

   Splitting data into multiple documents complicates queries and slows down updates.

4. Final Answer:

   Because storing all profile data in one document makes retrieval and updates simpler and faster -> Option A

5. Quick Check:

   One document per profile = simpler management [OK]

Hint: Keep related data in one document for speed [OK]

Common Mistakes:

• Thinking multiple documents per profile speed search
• Believing Elasticsearch limits one document per user
• Assuming documents must be very small always