Hi
@Thanhtan Le
Thanks for reaching out,
If I understand correctly, you are trying to query delta table directly (without using spark) and get the result as json records .
In that case , we can use delta-rs package or python variant of delta-rs package to read the delta table as arrow table (or pandas dataframe)and convert the result set into json records of required schema.
Delta table ——> delta-rs read —-> arrow table /pandas ——> json records
This project (lakehouse sharing) that you have mentioned is implemention of delta-sharing protocol which will give you json records of s3 uri of parquet files .that you need to download and read has pandas dataframe and send it as json .
Hope this helps!
Some Reference:
https://github.com/delta-io/delta-rs
https://roapi.github.io/docs/config/dataset-formats/delta.html