Support DynamoDBRead query operation (partition key / sort key filtering)

[nasudadada]

Is your feature request related to a problem? Please describe.
Currently, DynamoDBRead only supports the scan operation. This works but can be inefficient for large datasets and does not allow selective retrieval when partition key or sort key conditions are known. As a result, users may face performance issues and unnecessary read capacity consumption.

Describe the solution you'd like
Extend DynamoDBRead to support the query operation in addition to scan.

• Allow users to specify partition key and sort key conditions in the configuration.
• When these parameters are provided, use query instead of scan.
• Maintain backward compatibility: if no keys are specified, fall back to scan.

Describe alternatives you've considered

• Implement filtering on the client side after scan, but this is inefficient and costly for large tables.
• Creating a separate adapter only for query, but extending the current class would be more consistent.

Additional context
This was mentioned as a potential enhancement in PR:#458