Struct CscMatrix 

pub struct CscMatrix<T: Real = f64> {
    pub n_rows: usize,
    pub col_ptr: Vec<usize>,
    pub row_idx: Vec<u32>,
    pub val: Vec<T>,
}

Sparse matrix in compressed sparse column (CSC) format.

See the module documentation for the invariants. Equality is structural (same dimensions, same stored entries in the same order), which for matrices satisfying the invariants is equality of the matrices.

Fields

n_rows: usize

Number of rows m.

col_ptr: Vec<usize>

Column pointers, length n + 1; column j occupies col_ptr[j]..col_ptr[j + 1].

row_idx: Vec<u32>

Row index of each stored entry, strictly increasing within a column.

val: Vec<T>

Value of each stored entry; finite and nonzero.

Implementations

impl<T: Real> CscMatrix<T>

pub fn new(n_rows: usize, n_cols: usize) -> Self

The m × n zero matrix (the contract’s name; the same as CscMatrix::zeros).

Panics

If m or n exceeds u32::MAX.

pub fn zeros(n_rows: usize, n_cols: usize) -> Self

The m × n zero matrix.

Panics

If m or n exceeds u32::MAX.

pub fn from_raw_parts( n_rows: usize, col_ptr: Vec<usize>, row_idx: Vec<u32>, val: Vec<T>, ) -> Result<Self, SparseError>

Builds a matrix from raw compressed arrays, validating every invariant.

pub fn from_triplets( n_rows: usize, n_cols: usize, triplets: &[(Row, Col, T)], ) -> Result<Self, SparseError>

Builds an m × n matrix from triplets (i, j, v) by Algorithm C: duplicates are summed in input order and exact zeros are dropped.

pub fn n_rows(&self) -> usize

Number of rows m.

pub fn n_cols(&self) -> usize

Number of columns n.

pub fn nnz(&self) -> usize

Number of stored entries.

pub fn col(&self, j: Col) -> (&[u32], &[T])

Row indices and values of column j.

pub fn get(&self, i: Row, j: Col) -> T

Entry a_ij, or 0 if not stored; O(log nnz(column j)).

pub fn gaxpy(&self, alpha: T, x: &[T], y: &mut [T])

y ← y + α A x with dense x (length n) and y (length m), column by column: y_i ← y_i + a_ij (α x_j) for the stored entries, skipping the columns with x_j = 0.

Panics

If x.len() != n or y.len() != m.

pub fn to_csr(&self) -> CsrMatrix<T>

The same matrix in row-wise (CSR) storage, by Algorithm T.

pub fn transpose(&self) -> CscMatrix<T>

The transpose Aᵀ in CSC storage, by Algorithm T.

pub fn check_invariants(&self) -> Result<(), SparseError>

Checks invariants (C1)–(C4).

impl<T: Real> CscMatrix<T>

pub fn pattern(&self) -> Pattern

The structure without values, consumed by the orderings (Chapter 5.9).

pub fn symmetric_pattern(&self) -> Result<Pattern, SparseError>

The pattern of A + Aᵀ (Algorithm S) for the symmetric orderings; an error unless A is square.

Trait Implementations

impl<T: Clone + Real> Clone for CscMatrix<T>

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<T: Real> ColumnSource<T> for CscMatrix<T>

An explicit square matrix B.

fn m(&self) -> usize

Order m of B (its number of columns).

fn for_each_entry<F: FnMut(u32, T)>(&self, k: usize, f: F)

Calls f(i, b_ik) for every stored entry of column k, rows increasing for structural columns.

fn nnz_hint(&self) -> usize

Upper bound on the nonzeros of B (an allocation hint).

fn validate(&self) -> Result<(), SparseError>

Checks that the source describes a square m × m matrix whose references are in range (invariant (B1)).

impl<T: Debug + Real> Debug for CscMatrix<T>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<T: PartialEq + Real> PartialEq for CscMatrix<T>

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl<T: PartialEq + Real> StructuralPartialEq for CscMatrix<T>