GSI overloading technique in DynamoDB - Mini Project: Build & Apply

GSI Overloading Technique in DynamoDB

📋 What You'll Learn

💡 Why This Matters

🌍 Real World

This technique helps store multiple types of related data in a single DynamoDB table and query them efficiently using GSIs.

💼 Career

Understanding GSI overloading is important for designing scalable and cost-effective DynamoDB tables in real-world applications.

1

Create the DynamoDB table with primary keys

Create a DynamoDB table named UserActivities with UserId as the partition key and ActivityId as the sort key.

DynamoDB

# Create the DynamoDB table with UserId as partition key and ActivityId as sort key
# Your code here

Need a hint?

Define the table with UserId as the partition key and ActivityId as the sort key.

2

Add a Global Secondary Index (GSI) for activity type and timestamp

Add a Global Secondary Index named ActivityTypeTimestampIndex with ActivityType as the partition key and Timestamp as the sort key to the UserActivities table.

DynamoDB

CreateTable {
  TableName: "UserActivities",
  KeySchema: [
    { AttributeName: "UserId", KeyType: "HASH" },
    { AttributeName: "ActivityId", KeyType: "RANGE" }
  ],
  AttributeDefinitions: [
    { AttributeName: "UserId", AttributeType: "S" },
    { AttributeName: "ActivityId", AttributeType: "S" },
    { AttributeName: "ActivityType", AttributeType: "S" },
    { AttributeName: "Timestamp", AttributeType: "S" }
  ],
  GlobalSecondaryIndexes: [
    {
      IndexName: "ActivityTypeTimestampIndex",
      KeySchema: [
        { AttributeName: "ActivityType", KeyType: "HASH" },
        { AttributeName: "Timestamp", KeyType: "RANGE" }
      ],
      Projection: { ProjectionType: "ALL" }
    }
  ],
  BillingMode: "PAY_PER_REQUEST"
}
# Your code here

Need a hint?

Add ActivityType and Timestamp to AttributeDefinitions and define the GSI with these keys.

3

Insert sample items using GSI overloading technique

Insert sample items into UserActivities with different ActivityType values such as "login", "purchase", and "logout". Include UserId, ActivityId, ActivityType, and Timestamp attributes.

DynamoDB

CreateTable {
  TableName: "UserActivities",
  KeySchema: [
    { AttributeName: "UserId", KeyType: "HASH" },
    { AttributeName: "ActivityId", KeyType: "RANGE" }
  ],
  AttributeDefinitions: [
    { AttributeName: "UserId", AttributeType: "S" },
    { AttributeName: "ActivityId", AttributeType: "S" },
    { AttributeName: "ActivityType", AttributeType: "S" },
    { AttributeName: "Timestamp", AttributeType: "S" }
  ],
  GlobalSecondaryIndexes: [
    {
      IndexName: "ActivityTypeTimestampIndex",
      KeySchema: [
        { AttributeName: "ActivityType", KeyType: "HASH" },
        { AttributeName: "Timestamp", KeyType: "RANGE" }
      ],
      Projection: { ProjectionType: "ALL" }
    }
  ],
  BillingMode: "PAY_PER_REQUEST"
}

PutItem {
  TableName: "UserActivities",
  Item: {
    "UserId": { "S": "user123" },
    "ActivityId": { "S": "act1" },
    "ActivityType": { "S": "login" },
    "Timestamp": { "S": "2024-06-01T10:00:00Z" }
  }
}
# Your code here

Need a hint?

Insert multiple items with different ActivityType values and include UserId, ActivityId, and Timestamp.

4

Query activities by ActivityType using the GSI

Write a query to retrieve all items from UserActivities where ActivityType equals "login" using the ActivityTypeTimestampIndex GSI.

DynamoDB

CreateTable {
  TableName: "UserActivities",
  KeySchema: [
    { AttributeName: "UserId", KeyType: "HASH" },
    { AttributeName: "ActivityId", KeyType: "RANGE" }
  ],
  AttributeDefinitions: [
    { AttributeName: "UserId", AttributeType: "S" },
    { AttributeName: "ActivityId", AttributeType: "S" },
    { AttributeName: "ActivityType", AttributeType: "S" },
    { AttributeName: "Timestamp", AttributeType: "S" }
  ],
  GlobalSecondaryIndexes: [
    {
      IndexName: "ActivityTypeTimestampIndex",
      KeySchema: [
        { AttributeName: "ActivityType", KeyType: "HASH" },
        { AttributeName: "Timestamp", KeyType: "RANGE" }
      ],
      Projection: { ProjectionType: "ALL" }
    }
  ],
  BillingMode: "PAY_PER_REQUEST"
}

PutItem {
  TableName: "UserActivities",
  Item: {
    "UserId": { "S": "user123" },
    "ActivityId": { "S": "act1" },
    "ActivityType": { "S": "login" },
    "Timestamp": { "S": "2024-06-01T10:00:00Z" }
  }
}

PutItem {
  TableName: "UserActivities",
  Item: {
    "UserId": { "S": "user123" },
    "ActivityId": { "S": "act2" },
    "ActivityType": { "S": "purchase" },
    "Timestamp": { "S": "2024-06-01T10:05:00Z" }
  }
}

PutItem {
  TableName: "UserActivities",
  Item: {
    "UserId": { "S": "user123" },
    "ActivityId": { "S": "act3" },
    "ActivityType": { "S": "logout" },
    "Timestamp": { "S": "2024-06-01T10:10:00Z" }
  }
}

Query {
  TableName: "UserActivities",
  IndexName: "ActivityTypeTimestampIndex",
  KeyConditionExpression: "ActivityType = :atype",
  ExpressionAttributeValues: {
    ":atype": { "S": "login" }
  }
}
# Your code here

Need a hint?

Use the Query operation with IndexName set to ActivityTypeTimestampIndex and filter by ActivityType equals "login".