Day 10: Storing Financial Data in DynamoDB with Python (Boto3).

Welcome to Day 10. The "pipes" are connected. Now we need a bucket to catch the water.

Today, I designed the storage layer for my AI Financial Agent. I chose Amazon DynamoDB because it fits perfectly with the Serverless ethos (Lambda + EventBridge).

The Database Design (Single Table Concept)

Designing for NoSQL is different from SQL. You don't think about "joins", you think about "access patterns".

Query: "Give me all transactions for User Eric sorted by date."

PK: user_id

SK: timestamp

The Code

Here is how I updated my Lambda to save the mocked data:

Python
import boto3
import uuid
from datetime import datetime

dynamodb = boto3.resource('dynamodb')
table = dynamodb.Table('FinanceAgent-Transactions')

def save_transaction(description, amount):
item = {
'user_id': 'user_eric_01',
'transaction_date': datetime.now().isoformat(),
'tx_id': str(uuid.uuid4()),
'description': description,
'amount': amount,
'currency': 'EUR'
}

table.put_item(Item=item)
print(f"Saved: {description} - {amount}€")

Why On-Demand?

I configured the table as On-Demand Capacity. In a startup or learning phase, traffic is spiky or non-existent. On-Demand means AWS manages the read/write units automatically. I don't have to guess or provision servers.

Now my data is persistent. It survives the Lambda execution. Tomorrow: The AI integration begins.