`scipy.linalg._flapack:dgeqp3`

Summary

Compute a QR factorization with column pivoting (LAPACK dgeqp3).

a : ndarray: Matrix of shape (m, n).
lwork : int, optional: Size of the workspace. Default is 3 * (n + 1). Use the value reported in work[0] for the optimal value.
overwrite_a : int, optional: If nonzero, a may be overwritten in place. Default is 0.

qr : ndarray: R in the upper triangle; below it, the reflectors that define Q.
jpvt : ndarray: Column pivots, 1-based as LAPACK numbers them: column jpvt[i] - 1 of a became column i of a @ p.
tau : ndarray: Scalar factors of the elementary reflectors that define Q.
work : ndarray: Workspace actually used; work[0] is the optimal lwork.
info : int: 0 on success; if negative, the -info-th argument had an illegal value.