jaidisido

@Taragolis, first of all thank you ever so much for this feature request description. It's very clear and we appreciate the level of detail. Here are our thoughts on the...

Not yet - we still hope to implement a proper formatter across the library but must say it has slipped through other priorities

Contributions are always welcome @Taragolis, and we could indeed consider a breaking change for 3.0.0 but please let us know in which direction you would like to take this so...

> I will try to create built-in formatter just for avoid additional imports Agreed - that is most likely the right approach, perhaps you can take inspiration from the one...

Hi @maxispeicher, what makes you think that scan ranges don't make sense with Parquet? The [docs](https://docs.aws.amazon.com/AmazonS3/latest/userguide/selecting-content-from-objects.html) mention that it can be used for parquet: ``` Amazon S3 Select scan range...

Ah I see, did not realise you were using MAX in the query. So here is what happens: 1. the same query request ("SELECT MAX(c1)...") is made to the S3...

Yeah a good question I asked myself as well. I don't know the internals of the S3 select engine to be honest, but my guess would be that the MAX...

True, it would indeed work if we make `use_threads=False` work on an entire file instead of multiple scan ranges. The only concern I have at that point is how unintuitive...

Any reason why you are not simply passing an S3 prefix instead of a list of objects? With an S3 prefix argument (e.g. `s3://my-bucket/my-prefix/`) the method lists all objects, ensuring...

> Looks like some cyclic imports there No there are no cyclic imports, I can run pylint without issue locally and I can import the library too. Plus, no change...