Module selection 

Logic for selecting which rows to read: RowSelection and RowSelector

This module holds RowSelection and its public API, which dispatches to one of the two backings depending on how the selection is stored:

• selector: the run length backing, RowSelector and its primitives
• boolean: the bitmap backing, MaskSelection and its primitives

The remaining modules hold the operations that are common to both:

• algebra: and_then, intersection and union
• ranges: mapping a RowSelection onto page and batch ranges
• cursor: iterating a RowSelection while reading

Re-exports

use crate::file::page_index::offset_index::PageLocation; 🔒

use arrow_array::Array; 🔒

use arrow_array::BooleanArray; 🔒

use arrow_buffer::BooleanBuffer; 🔒

use arrow_buffer::BooleanBufferBuilder; 🔒

use arrow_select::filter::SlicesIterator; 🔒

use std::cmp::Ordering; 🔒

use std::collections::VecDeque; 🔒

use std::ops::Range; 🔒

use algebra::and_then_mask; 🔒

use algebra::and_then_row_selections; 🔒

use algebra::and_then_selectors_with_mask; 🔒

use algebra::combine_mixed_selection; 🔒

use algebra::intersect_masks; 🔒

use algebra::intersect_row_selections; 🔒

use algebra::union_masks; 🔒

use algebra::union_row_selections; 🔒

pub use boolean::MaskRunIter;

use boolean::MaskSelection; 🔒

use boolean::boolean_mask_from_selectors; 🔒

use boolean::limit_mask; 🔒

use boolean::mask_has_at_least_runs; 🔒

use boolean::offset_mask; 🔒

use boolean::split_off_mask; 🔒

use boolean::trim_mask; 🔒

pub(crate) use cursor::LoadedRowRanges; 🔒

pub(crate) use cursor::MaskCursor; 🔒

pub(crate) use cursor::RowSelectionStrategy; 🔒

pub use cursor::RowSelectionCursor;

pub use cursor::RowSelectionPolicy;

use ranges::expand_to_batch_boundaries_from_selectors; 🔒

use ranges::scan_ranges_from_mask; 🔒

use ranges::scan_ranges_from_selectors; 🔒

pub use selector::RowSelector;

use selector::limit_selectors; 🔒

use selector::offset_selectors; 🔒

use selector::split_off_selectors; 🔒

Modules

algebra 🔒

Set algebra backing RowSelection::and_then, RowSelection::intersection and RowSelection::union

boolean 🔒

The bitmap backed representation of a RowSelection and the primitives operating on it: conversion to and from the run length (RowSelector) form, and the transforms backing split_off, trim, offset and limit.

cursor 🔒

Execution time iteration over a RowSelection.

ranges 🔒

Mapping row selections onto ranges: the byte ranges of the data pages that must be fetched (RowSelection::scan_ranges) and the expansion of a selection to batch boundaries.

selector 🔒

The run length backed representation of a RowSelection: RowSelector and the primitives operating on a sequence of them.

Structs

RowSelection

RowSelection represents selecting a subset of rows when scanning a parquet file.

Enums

RowSelectionInner 🔒

Internal storage for RowSelection.

Functions

append_normalized_selector 🔒

Appends a selector while maintaining the normalized selector invariants.

auto_min_mask_runs 🔒

Returns the minimum normalized run count for which the Auto selection policy would prefer a mask to RLE.

filters_to_boolean_buffer 🔒

selectors_below_run_limit 🔒

Builds normalized selectors while their run count remains below min_mask_runs. Returning None drops the partial selector allocation before the caller constructs a mask.