[Parquet] Improve speed of conversion between RowSelection <--> Mask

[alamb]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.

After the great work from @hhhizzz in #8733, we will (finally) have the ability to use a Bitmask filter representation when applying filters during Parquet decode.

During review, one thing we noticed is that the code that is used to convert a RowSelection to a Mask could likely be optimized more

arrow-rs/parquet/src/arrow/arrow_reader/selection.rs

Lines 926 to 933 in 911331a

	fn boolean_mask_from_selectors(selectors: &[RowSelector]) -> BooleanBuffer {
	let total_rows: usize = selectors.iter().map(|s| s.row_count).sum();
	let mut builder = BooleanBufferBuilder::new(total_rows);
	for selector in selectors {
	builder.append_n(selector.row_count, !selector.skip);
	}
	builder.finish()
	}

Describe the solution you'd like

Make predicate evaluation faster by optimizing the conversion to mask

Describe alternatives you've considered
@hhhizzz mentions on #8733 (comment)

I'll learn the #6624 (comment) and try to improve it in the following PRs.

Aka that @XiangpengHao 's PR here has many tricks to use

• [Parquet] Add BooleanArray based row selection #6624

Additional context

[sdf-jkl]

It seems like the #8733 implementation is the same as one in #6624.
Impls respectively:

fn boolean_mask_from_selectors(selectors: &[RowSelector]) -> BooleanBuffer {
    let total_rows: usize = selectors.iter().map(|s| s.row_count).sum();
    let mut builder = BooleanBufferBuilder::new(total_rows);
    for selector in selectors {
        builder.append_n(selector.row_count, !selector.skip);
    }
    builder.finish()
}

impl From<RowSelection> for BooleanRowSelection {
    fn from(selection: RowSelection) -> Self {
        let total_rows = selection.row_count();
        let mut builder = BooleanBufferBuilder::new(total_rows);

        for selector in selection.iter() {
            if selector.skip {
                builder.append_n(selector.row_count, false);
            } else {
                builder.append_n(selector.row_count, true);
            }
        }

        BooleanRowSelection {
            selector: builder.finish(),
        }
    }
}

[alamb]

So do you suggest we close this issue?

[sdf-jkl]

I played around with codex a bit last week and it did suggest some improvements, but I haven't benchmarked them yet.

[sdf-jkl]

@alamb I think we can close this. The way we use BooleanBufferBuilder here is clean and any extra optimizations would require changing the BooleanBufferBuilder API, which is a different issue. (I don't see any optimization there either)

[alamb]

Thanks @sdf-jkl -- done