Backend interview prep

Start

DynamoDB GSI vs LSI Interview Questions

Master DynamoDB Secondary Indexes with interview-focused questions covering Global Secondary Index (GSI), Local Secondary Index (LSI), sparse indexes, projections, query optimization, performance, costs, and production best practices.

Introduction

One of the biggest limitations of DynamoDB is that you cannot efficiently query by arbitrary attributes.

Unlike relational databases where indexes can be added on almost any column, DynamoDB requires you to carefully design indexes around your application's access patterns.

DynamoDB provides two types of secondary indexes:

  • Global Secondary Index (GSI)
  • Local Secondary Index (LSI)

Choosing the correct index type significantly impacts:

  • Query Performance
  • Cost
  • Scalability
  • Storage
  • Throughput

This chapter covers the most frequently asked DynamoDB indexing interview questions.

Secondary Index Architecture

flowchart LR

Application --> PrimaryTable

PrimaryTable --> GSI

PrimaryTable --> LSI

GSI --> Query

LSI --> Query

1. What is a Secondary Index?

Answer

A Secondary Index provides an alternative way to query data without scanning the entire table.

Benefits

  • Faster Queries
  • Multiple Access Patterns
  • Better Performance
  • Lower Cost than Scan

2. Why are Secondary Indexes required?

Suppose a table uses

Partition Key

CustomerId

Now the application needs

Find Customer By Email

Without an index

↓

Full Table Scan

With an index

↓

Efficient Query

3. What are the two types of Secondary Indexes?

DynamoDB supports

  • Global Secondary Index (GSI)
  • Local Secondary Index (LSI)

4. What is a Global Secondary Index (GSI)?

A GSI is an index with its own Partition Key and optional Sort Key.

It enables completely different query patterns from the base table.

GSI Architecture

flowchart LR

PrimaryTable --> GSI

GSI --> DifferentPartitionKey --> Query

5. What is a Local Secondary Index (LSI)?

An LSI uses

  • Same Partition Key
  • Different Sort Key

It allows additional sorting within the same partition.

LSI Architecture

flowchart LR

PrimaryTable --> SamePartitionKey --> DifferentSortKey --> Query

6. Difference between GSI and LSI?

GSI LSI
Different Partition Key Same Partition Key
Optional Sort Key Different Sort Key
Created Anytime Must be Created with Table
Separate Throughput Shares Table Throughput
Eventually Consistent Strong or Eventual Reads

7. When should GSI be used?

Use GSI when

  • New Query Pattern
  • Different Partition Key
  • Search by Email
  • Search by Status
  • Search by Category

Example

Table

PK = CustomerId

Need

Find by Email

Create

GSI

PK = Email

8. When should LSI be used?

Use LSI when

Partition Key remains the same

but sorting changes.

Example

PK = CustomerId

Sort by

OrderDate

or

OrderAmount

9. Can GSI have a different Partition Key?

Yes.

Example

Primary Table

CustomerId

GSI

Email

10. Can LSI have a different Partition Key?

No.

LSI always shares the same Partition Key.

11. Can GSI have a different Sort Key?

Yes.

Example

PK

Email

SK

CreatedDate

12. Can LSI have a different Sort Key?

Yes.

Changing the Sort Key is the main purpose of LSI.

13. Can GSI be created after table creation?

Yes.

GSI can be

  • Added
  • Modified
  • Deleted

without recreating the table.

14. Can LSI be created later?

No.

LSI must be created when the table is created.

15. Which index has separate throughput?

GSI

has its own

  • RCU
  • WCU

16. Which index shares throughput?

LSI shares throughput with the base table.

17. Which index supports Strongly Consistent Reads?

LSI

supports

  • Strong Consistency
  • Eventual Consistency

18. Which index only supports Eventually Consistent Reads?

Global Secondary Index

19. What is an Index Projection?

Projection determines

which attributes are copied into an index.

Types

  • KEYS_ONLY
  • INCLUDE
  • ALL

Projection Types

flowchart LR

Table --> Projection

Projection --> KEYS_ONLY

Projection --> INCLUDE

Projection --> ALL

20. Explain KEYS_ONLY Projection.

Stores only

  • Primary Keys
  • Index Keys

Lowest storage cost.

21. Explain INCLUDE Projection.

Stores

Keys

Selected Attributes

Balances cost and performance.

22. Explain ALL Projection.

Copies every attribute.

Advantages

  • Fast Queries

Disadvantages

  • Higher Storage Cost

23. What is a Sparse Index?

A Sparse Index contains only items that include the indexed attribute.

Example

Only active orders have

Status

Only those items appear in the index.

Sparse Index Example

Orders

↓

Status Exists

↓

Indexed

↓

Query

24. Does every item appear in a GSI?

No.

Only items containing the indexed attribute.

25. How does GSI improve performance?

Instead of

Scan

↓

Entire Table

Use

Query

↓

Index

26. Does GSI increase write cost?

Yes.

Every write updates

  • Base Table
  • Index

Multiple GSIs increase write cost.

27. Does LSI increase storage?

Yes.

Additional index storage is required.

28. Banking Example

Primary Table

PK

AccountId

SK

TransactionTime

Need

Find Transactions

By Status

Create

GSI

PK

Status

SK

TransactionTime

29. E-Commerce Example

Primary

CustomerId

OrderDate

Need

Orders By Product

Create

GSI

ProductId

OrderDate

30. HR Example

Primary

EmployeeId

Need

Employees By Department

Create

GSI

Department

EmployeeName

31. Time-Series Example

Primary

DeviceId

Timestamp

Need

Events By Severity

Create

GSI

Severity

Timestamp

32. When should Scan be replaced with GSI?

Whenever the application repeatedly searches by a non-key attribute.

Example

Instead of

Scan

WHERE Status='OPEN'

Use

GSI

PK

Status

33. Common mistakes

  • Too many GSIs
  • Wrong Partition Key
  • Using Scan instead of Query
  • Large ALL projections
  • Creating unnecessary indexes
  • Ignoring write costs

Enterprise Best Practices

  • Design indexes around access patterns.
  • Prefer Query over Scan.
  • Keep GSI count minimal.
  • Choose projection type carefully.
  • Use KEYS_ONLY whenever possible.
  • Use INCLUDE for frequently accessed attributes.
  • Monitor GSI throttling.
  • Review index usage regularly.
  • Avoid indexing low-value attributes.
  • Consider write amplification before adding indexes.

Quick Revision

Feature GSI LSI
Partition Key Different Same
Sort Key Optional Different
Create Later Yes No
Delete Later Yes No
Strong Reads No Yes
Separate Throughput Yes No
Storage Additional Additional
Best For New Access Patterns Alternate Sorting

Interview Tips

Interviewers commonly ask

  • Difference between GSI and LSI.
  • Why use GSI instead of Scan?
  • Can GSI be added later?
  • Why can't LSI be added later?
  • Explain Sparse Index.
  • Explain Projection Types.
  • Difference between KEYS_ONLY and ALL.
  • Which index supports Strongly Consistent Reads?
  • Does GSI increase write cost?
  • Give a banking use case for GSI.

Always explain the trade-off between query flexibility, write cost, storage cost, and consistency when discussing secondary indexes.

Summary

Global Secondary Indexes (GSIs) and Local Secondary Indexes (LSIs) enable DynamoDB to support multiple access patterns without performing expensive table scans. GSIs provide new partition keys and independent scalability, while LSIs reuse the existing partition key with alternative sorting.

Understanding secondary indexes, projection types, sparse indexes, throughput implications, and consistency behavior is essential for designing scalable DynamoDB applications and succeeding in AWS, backend engineering, cloud architecture, and solution architect interviews.