Array slicing

Array slicing

Programming One-Dimensional Finite Element Method for the Linear Boundary Value Problem

Published in Pavel Sumets, Computational Framework for the Finite Element Method in MATLAB® and Python, 2023

Array slicing and indexing works exactly the same way it does for lists except that they can be applied to multiple dimensions as well. For example, given a matrix A, the command A[0,3] accesses a single element of the matrix, and A[:,3] extracts a column. Note that slices of arrays do not copy the internal array data but only produce new views of the original data, so if b=A[0,3] and b=b+1 then element of matrix A would be modified implicitly as well. If a copy is needed, then an explicit copy() command is required. NumPy arrays may be indexed with other arrays or lists. For all cases of index arrays, the copy of original data is returned, not a view. For example A[np.array([0,2,4]),0] or A[np.array([0,2,4])][0] returns elements of the first column with indices 0, 2 and 4.

p†q: a tool for prototyping many-body methods for quantum chemistry

View Article

Journal Information

Published in Molecular Physics, 2021

Nicholas C. Rubin, A. Eugene DePrince

An important part of computer algebra systems is the potential to translate equations into usable code. pq comes with a front-end parser that produces code for residual equations in a NumPy einsum convention. Appropriate summation limits, occupied or virtual indices, are enforced using NumPy array slicing. Using the NumPy einsum format allows a user to implement the generated code with any of a variety of einsum backends. For example, tensor contraction engines in Tensorflow [97], Jax [98], or Dask [99] can be used to implement coupled cluster iterations on a variety of different hardware platforms.

Explore chapters and articles related to this topic

Programming One-Dimensional Finite Element Method for the Linear Boundary Value Problem

p†q: a tool for prototyping many-body methods for quantum chemistry