Fast and robust slice-based algorithms (e.g., [sorting], [selection], [search]) for
non-contiguous (sub)views into n-dimensional arrays. Reimplements algorithms in [slice] for
[ndarray] with arbitrary memory layout (e.g., non-contiguous).
```rust use ndarray_slice::{Slice1Ext, ndarray::arr2};
// 2-dimensional array of 4 rows and 5 columns. let mut v = arr2(&[[-5, 4, 1, -3, 2], // row 0, axis 0 [ 8, 3, 2, 4, 8], // row 1, axis 0 [38, 9, 3, 0, 3], // row 2, axis 0 [ 4, 9, 0, 8, -1]]); // row 3, axis 0 // \ \ \ // column 0 \ column 4 axis 1 // column 2 axis 1
// Mutable subview into the last column. let mut column = v.column_mut(4);
// Due to row-major memory layout, columns are non-contiguous and hence cannot be sorted by // viewing them as mutable slices. asserteq!(column.asslice_mut(), None);
// Instead, sorting is specifically implemented for non-contiguous mutable (sub)views. column.sort();
assert!(v == arr2(&[[-5, 4, 1, -3, -1], [ 8, 3, 2, 4, 2], [38, 9, 3, 0, 3], [ 4, 9, 0, 8, 8]])); // \ // column 4 sorted, others untouched ```
Complexities where n is the length of the (sub)view.
| Algorithm | Stable | Allocation | Recursive | Average | Worse-Case | |-----------|------- |------------|-----------|----------|-------------------| | Sorting | yes | yes | no | O(n) | O(n log(n)) | | Sorting | no | no | yes | O(n) | O(n log(n)) | | Selection | no | no | no | O(n) | O(n log(n)) |
SliceExt trait for n-dimensional array or (sub)view with methods expecting Axis as
their first argument. Comparing methods will always be suffixed with _by or _by_key
defining how to compare multi-dimensional elements (e.g., columns) along the provided axis
of interest (e.g., rows).See the release history to keep track of the development.
alloc: Enables stable sort. Enabled by std.std: Enables selection of many indices. Enabled by default.Copyright © 2023 Rouven Spreckels rs@qu1x.dev
This project is licensed under either of
at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this project by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.