Data loaders for in-memory modern OLAP tools (DuckDB and Polars)

[JohnLyu94]

Hello, this is my initial idea after roughly browsing CAMEL's modules. Since I have just started to get involved in this community, I would like to ask in the discussion before submitting an issue.

I found panda_reader in Data Loaders. Ideally, there could be similar loaders for more recent in-memory OLAP libraries such as DuckDB and Polars.

There are a few potential benefits. 1) They can provide similar functionality to pandas and are much more efficient; 2) They are better API compatible with modern columnar table format and cloud storage; 3) These new OLAP libraries make extensive use of Apache Arrow as the data format behind the scenes, providing strong type support and the ability to schedule data between processes. Overall, introducing them may help us enhance the data-related functionalities.