DynamoDB: What is it?

Amazon Web Services (AWS) offers a NoSQL database service called DynamoDB.

• It adheres to the ACID (Atomicity, Consistency, Isolation, and Durability) properties.

• It can process millions of requests per second and is very scalable.

Important characteristics:

• Ready for production from the first day

• Pay-as-you-go pricing structure

• Infrastructure that is automatically scalable

• Smooth interaction with Amazon S3

• It is renowned for being easy to use and integrate with other services

• The official AWS documentation is the best source for in-depth DynamoDB knowledge

---

DynamoDB's Primary Key

The Primary Key for each DynamoDB table can be:

• A basic Partition Key, or

• A composite key: Partition Key + Sort Key

Each item must have a distinct partition key and sort key combination.

---

DynamoDB Secondary Indexes

To improve data retrieval, DynamoDB offers two kinds of secondary indexes:

1. The Global Secondary Index (GSI)

• May differ from the primary table in terms of the Sort Key and Partition Key

• Can be added or removed at any time, even once the table is operational

• Up to 20 GSIs are supported by each table

• Beneficial for alternative access routes and adaptable query patterns

2. The LSI, or Local Secondary Index

• Must have a different Sort Key but the same Partition Key as the main table

• Cannot be removed after the table has been created

• Up to five LSIs are supported by each table

• Helpful for sorting the same partition key according to various criteria

---

Why Make Use of LSIs and GSIs?

• For certain query patterns, the default key structure may not be the best option

• Secondary indexes facilitate more effective data querying

• They enhance performance and steer clear of full-table scans

• But it's important to pick the right keys and indexes early on

---

Example Table: Scores for Games

Assume that we are constructing a table for a gaming platform.
This is the structure of our base table:

---

Case 1: Sorting User Scores with LSI

Table's primary key:

• user_id is the Partition Key

• game_id is the Sort Key

Issue:

All of the games that user U1 has played should be arranged according to their highest score.
After fetching, you would need to manually sort using the base table.

Answer:

Make an LSI using:

• user_id as the Partition Key (same as table)

• Sort Key: highest_score

You will now receive this sorted result after retrieving the U1 scores:

Because the LSI does the sorting internally, this query is now quick and effective.

---

Case 2: Querying Users by Game with GSI

All users who have played game G1 should be retrieved, arranged according to the number of times they have played.

Issue:

Since game_id is not a Partition Key in the base table, the query would need to scan every record.

Answer:

Make a GSI using:

• game_id as the Partition Key

• games_played as the Sort Key

Now, the GSI provides you with the following when you query for game_id = G1:

When sorted by games_played, this is:

A lot quicker than going through the entire table.

---

Important Takeaways

• DynamoDB is a robust NoSQL service that has excellent performance and scalability built in

• Optimized query patterns are made possible by secondary indexes (GSI and LSI)

• While LSIs must be created initially, GSIs are adaptable and can be added at any time

• If you want different sorting logic for the same Partition Key, use LSI

• When you need to query using a completely different attribute, use GSI

• Performance depends on selecting the appropriate keys and indexes according to anticipated access patterns

---

Last Remark

After reading the official AWS documentation for the entire day, these were my last thoughts regarding GSI, LSI, and DynamoDB. If you want to learn more, I suggest reading the official AWS documents.