Lambda trigger on stream events in DynamoDB

Introduction

Lambda triggers on stream events let you run code automatically when your database changes. This helps you react quickly to updates without checking manually.

You want to send a notification when a new item is added to your database.

You need to update another system whenever data changes in your table.

You want to keep a backup of changes made to your database.

You want to analyze data changes in real-time for reports.

You want to clean or validate data right after it is inserted or updated.

Syntax

DynamoDB

1. Enable DynamoDB Streams on your table.
2. Create a Lambda function.
3. Configure the Lambda function as a trigger for the DynamoDB stream.

Example JSON event structure received by Lambda:
{
  "Records": [
    {
      "eventID": "string",
      "eventName": "INSERT | MODIFY | REMOVE",
      "dynamodb": {
        "Keys": { ... },
        "NewImage": { ... },
        "OldImage": { ... },
        "SequenceNumber": "string",
        "SizeBytes": number,
        "StreamViewType": "NEW_IMAGE | OLD_IMAGE | NEW_AND_OLD_IMAGES | KEYS_ONLY"
      },
      "eventSource": "aws:dynamodb",
      "eventVersion": "1.1",
      "eventSourceARN": "string"
    }
  ]
}

You must enable DynamoDB Streams on your table before Lambda can receive events.

The Lambda function receives an event object containing records of changes.

Examples

This Lambda function logs the type of change and the new data for each record.

DynamoDB

exports.handler = async (event) => {
  for (const record of event.Records) {
    console.log('Event Name:', record.eventName);
    console.log('New Image:', JSON.stringify(record.dynamodb.NewImage));
  }
};

This function only processes new items added to the table.

DynamoDB

exports.handler = async (event) => {
  event.Records.forEach(record => {
    if (record.eventName === 'INSERT') {
      console.log('New item added:', JSON.stringify(record.dynamodb.NewImage));
    }
  });
};

Sample Program

This Lambda function collects all new items added in the event and logs them together.

DynamoDB

exports.handler = async (event) => {
  const addedItems = [];
  for (const record of event.Records) {
    if (record.eventName === 'INSERT') {
      addedItems.push(record.dynamodb.NewImage);
    }
  }
  console.log('Added items:', JSON.stringify(addedItems));
};

Important Notes

Lambda triggers run automatically and do not require manual polling.

StreamViewType setting controls what data is sent to Lambda (new image, old image, or both).

Remember to give your Lambda function permission to read from the DynamoDB stream.

Summary

Lambda triggers on DynamoDB streams let you run code when data changes.

Enable streams, create a Lambda, and connect them as a trigger.

Use the event data to react to inserts, updates, or deletes automatically.

Practice

1. What is the main purpose of a Lambda trigger on DynamoDB stream events?

easy

A. To backup the DynamoDB table data to S3

B. To manually update the DynamoDB table from the Lambda console

C. To disable the DynamoDB stream temporarily

D. To automatically run code when data in the DynamoDB table changes

Solution

1. Step 1: Understand DynamoDB streams

   DynamoDB streams capture data changes like inserts, updates, and deletes.

2. Step 2: Role of Lambda triggers

   Lambda triggers run code automatically when these stream events occur.

3. Final Answer:

   To automatically run code when data in the DynamoDB table changes -> Option D

4. Quick Check:

   Lambda triggers = auto-run code on data change [OK]

Hint: Lambda triggers run code automatically on data changes [OK]

Common Mistakes:

• Thinking Lambda triggers require manual execution
• Confusing stream enabling with backup
• Assuming Lambda disables streams

2. Which of the following is the correct way to specify a Lambda trigger for a DynamoDB stream in AWS Console?

easy

A. Enable stream on the table, then add Lambda as a trigger on the stream

B. Create Lambda function first, then enable stream on the table

C. Enable stream on the table and Lambda separately without linking

D. Add Lambda trigger before enabling the stream on the table

Solution

1. Step 1: Enable stream on DynamoDB table

   Streams must be enabled first to capture data changes.

2. Step 2: Add Lambda as a trigger on the enabled stream

   Lambda triggers are linked to the stream to react to events.

3. Final Answer:

   Enable stream on the table, then add Lambda as a trigger on the stream -> Option A

4. Quick Check:

   Enable stream first, then add Lambda trigger [OK]

Hint: Enable stream first, then attach Lambda trigger [OK]

Common Mistakes:

• Trying to add Lambda trigger before enabling stream
• Not linking Lambda to the stream
• Enabling stream and Lambda separately without connection

3. Given this Lambda function snippet triggered by a DynamoDB stream event:

def lambda_handler(event, context):
    for record in event['Records']:
        if record['eventName'] == 'INSERT':
            print(record['dynamodb']['NewImage']['id']['S'])

What will this code output when a new item with id '123' is inserted?

medium

A. It prints '123' to the logs

B. It prints the old image of the item

C. It raises a KeyError because 'NewImage' is missing

D. It prints nothing because eventName is 'MODIFY'

Solution

1. Step 1: Check eventName condition

   The code checks if eventName is 'INSERT', which matches new item insertion.

2. Step 2: Access NewImage id value

   It accesses 'NewImage' dictionary and prints the 'id' string value '123'.

3. Final Answer:

   It prints '123' to the logs -> Option A

4. Quick Check:

   INSERT event prints new id '123' [OK]

Hint: INSERT event uses NewImage to get new item data [OK]

Common Mistakes:

• Confusing NewImage with OldImage
• Assuming eventName is MODIFY instead of INSERT
• Not accessing nested dictionary keys correctly

4. You wrote this Lambda code to process DynamoDB stream events but it throws a runtime error:

def lambda_handler(event, context):
    for record in event['Records']:
        print(record['dynamodb']['NewImage']['id']['S'])

What is the likely cause of the error?

medium

A. The 'id' attribute is always an integer, not a string

B. The event may contain DELETE events without NewImage data

C. The event['Records'] key is missing

D. Lambda functions cannot access DynamoDB streams

Solution

1. Step 1: Understand stream event types

   DELETE events do not have 'NewImage', only 'OldImage'.

2. Step 2: Accessing NewImage without check causes error

   Trying to access 'NewImage' on DELETE events causes a KeyError.

3. Final Answer:

   The event may contain DELETE events without NewImage data -> Option B

4. Quick Check:

   DELETE events lack NewImage causing runtime error [OK]

Hint: Check event type before accessing NewImage [OK]

Common Mistakes:

• Assuming all events have NewImage
• Ignoring DELETE event structure
• Believing Lambda can't access streams

5. You want your Lambda triggered by DynamoDB streams to process only UPDATE events where the attribute 'status' changed to 'completed'. Which approach is best?

hard

A. Filter events in DynamoDB stream settings to only send UPDATE events with 'completed' status

B. Use a separate Lambda for each event type and status value

C. In Lambda, check if eventName is 'MODIFY' and compare OldImage and NewImage 'status' values

D. Enable stream with only KEYS_ONLY to reduce data size

Solution

1. Step 1: Understand stream filtering limits

   DynamoDB streams cannot filter by attribute values, only event types.

2. Step 2: Use Lambda code to check eventName and compare status

   Lambda can check if eventName is 'MODIFY' and compare OldImage and NewImage 'status' to detect change to 'completed'.

3. Final Answer:

   In Lambda, check if eventName is 'MODIFY' and compare OldImage and NewImage 'status' values -> Option C

4. Quick Check:

   Filter in Lambda by event and attribute change [OK]

Hint: Filter attribute changes inside Lambda, not in stream settings [OK]

Common Mistakes:

• Trying to filter by attribute in stream settings
• Using KEYS_ONLY which lacks attribute data
• Creating multiple Lambdas unnecessarily