.reshape()

The .reshape() function rearranges the data in an ndarray into a new shape. The new shape must be compatible with the old one, though an index of -1 can be used to infer one dimension.

Syntax

numpy.reshape(array, newshape, order = 'C')

• array: The input array to be reshaped.
• newshape: An integer or a tuple representing the desired shape of the new array. If one dimension is set to -1, that dimension will be inferred based on the size of the original array.
• order: Specifies how elements should be read from the original array and placed into the reshaped array. It can be set to 'C', 'F', or 'A':
  • C: Read/write elements in row-major order (C-like), where the last axis (columns) changes fastest, and the first axis (rows) changes slowest. The elements are placed row by row.
  • F: Read/write elements in column-major order (Fortran-like), where the first axis (rows) changes fastest, and the last axis (columns) changes slowest. The elements are placed column by column.
  • A: Use Fortran-like index order if the array is Fortran contiguous in memory (i.e., stored sequentially without gaps), or C-like order otherwise.

Using 'C' or 'F' in .reshape() affects the indexing order but does not alter the physical layout of the data in memory.

If possible, the ndarray returned will be a view of the data from the original ndarray.

Example

The following example creates an ndarray then uses .reshape() to change its dimensions:

import numpy as np

nd1 = np.array([[1,2,3],[4,5,6]])

print(nd1)
print(np.reshape(nd1,(3,2)))
print(np.reshape(nd1,(-1,1)))

This produces the following output:

[[1 2 3]
 [4 5 6]]

[[1 2]
 [3 4]
 [5 6]]

[[1]
 [2]
 [3]
 [4]
 [5]
 [6]]

Codebyte Example

The following example creates a ndarray and then uses order as an optional parameter for .reshape() to change its dimensions:

us
Visit us
code
Hide code
Code
Output
Hide output
Hide output
Loading...
Run