DynamoDB is inherently structured around hash-based partitions. Because of this, it is not designed to support queries like "fetch all products starting with the letter D sorted alphabetically." Attempting to do this with a standard scan is extremely slow and expensive.

If your storefront app requires an A-Z directory index for SEO purposes, you must structure your data models specifically to support this query pattern.

1. Composite Partition-Sort Keys

To group items alphabetically, we create a DynamoDB table with a Partition Key (PK) storing the first letter of the product name (e.g., letter#D) and a Sort Key (SK) storing the full product name. Since DynamoDB stores items within a partition sorted by their sort key, we can query all items starting with D in perfect alphabetical order:

# Querying alphabetical lists in Python
def get_parts_by_letter(table_name, letter, limit=50, start_key=None):
    params = {
        "TableName": table_name,
        "KeyConditionExpression": "PK = :pk",
        "ExpressionAttributeValues": {":pk": {"S": f"letter#{letter.upper()}"}},
        "Limit": limit
    }
    if start_key:
        params["ExclusiveStartKey"] = start_key
        
    return dynamodb.query(**params)

2. The Available Letters Table

To avoid querying empty partitions (e.g., searching for products starting with X when none exist), we maintain a secondary, lightweight lookup table goseanto-available-part-number-first-letters. The storefront reads this table to render active links for letters, preventing dead-ends for users.