Struct RowSelection

Source

pub struct RowSelection { /* private fields */ }

Expand description

RowSelection allows selecting or skipping rows when scanning an ORC file.

This is applied prior to reading column data, and can therefore be used to skip IO to fetch data into memory, improving query performance.

A typical use-case would be using ORC stripe statistics or file-level indexes to filter out rows that don’t satisfy a predicate.

§Example

use orc_rust::row_selection::{RowSelection, RowSelector};

// Create selectors: skip 100 rows, select 50, skip 200
let selectors = vec![
    RowSelector::skip(100),
    RowSelector::select(50),
    RowSelector::skip(200),
];

let selection: RowSelection = selectors.into();

// Query properties
assert_eq!(selection.row_count(), 350);
assert_eq!(selection.selects_any(), true);

A RowSelection maintains the following invariants:

It contains no RowSelector with 0 rows
Consecutive RowSelectors alternate between skipping and selecting rows

Implementations§

Source §

impl RowSelection

Source

pub fn new() -> Self

Create a new empty RowSelection

Source

pub fn from_filters(filters: &[BooleanArray]) -> Self

Create a RowSelection from a slice of BooleanArray

§Panics

Panics if any of the BooleanArray contain nulls

Source

pub fn from_consecutive_ranges<I: Iterator<Item = Range<usize>>>( ranges: I, total_rows: usize, ) -> Self

Create a RowSelection from an iterator of consecutive ranges to keep

§Arguments

ranges - Iterator of consecutive ranges (e.g., 10..20, 30..40)
total_rows - Total number of rows in the stripe/file

§Example

use orc_rust::row_selection::RowSelection;

// Select rows 10-19 and 30-39 out of 50 total rows
let selection = RowSelection::from_consecutive_ranges(
    vec![10..20, 30..40].into_iter(),
    50
);

Source

pub fn select_all(row_count: usize) -> Self

Create a RowSelection that selects all row_count rows

Source

pub fn skip_all(row_count: usize) -> Self

Create a RowSelection that skips all row_count rows

Source

pub fn row_count(&self) -> usize

Returns the total number of rows (selected + skipped)

Source

pub fn selected_row_count(&self) -> usize

Returns the number of selected rows

Source

pub fn skipped_row_count(&self) -> usize

Returns the number of skipped rows

Source

pub fn selects_any(&self) -> bool

Returns true if this selection selects any rows

Source

pub fn iter(&self) -> impl Iterator<Item = &RowSelector>

Returns an iterator over the RowSelectors

Source

pub fn selectors(&self) -> &[RowSelector]

Returns a slice of the underlying RowSelectors

Source

pub fn split_off(&mut self, row_count: usize) -> Self

Splits off the first row_count rows from this RowSelection

Returns a new RowSelection containing the first row_count rows, and updates self to contain the remaining rows.

§Example

use orc_rust::row_selection::{RowSelection, RowSelector};

let mut selection = RowSelection::from_consecutive_ranges(
    vec![10..20, 30..40].into_iter(),
    50
);

let first = selection.split_off(25);
assert_eq!(first.row_count(), 25);
assert_eq!(selection.row_count(), 25);