add note about arrow::open_dataset

[eitsupi]

Hi, thank you for creating this wonderful book!

I have two concerns about the "From a parquet file" section, which can be read as a comparison with the arrow package, and propose the following corrections.

• Use as_data_frame = FALSE in arrow::read_parquet(). If this is not used, Acero is not used and just an R DataFrame calculation is performed.
• Mention to arrow::open_dataset(). I think there is a risk of misleading people into thinking that the lazy scan functionality does not exist in the arrow package.

[ddotta]

Thanks @eitsupi for this suggestion!

[eitsupi]

Thanks for merging and re rendering!