How to vectorize Numpy Array Modification like in Matlab

Question

Using Matlab you can modify an array from indices using vectorization :

A = [1, 2, 3, 4, 5, 6, 7, 8, 9]

Output :

A =

 1     2     3     4     5     6     7     8     9

Modifying :

A([2,4,6:9]) = -1

Output :

A =

 1    -1     3    -1     5    -1    -1    -1    -1

How can we do that using Numpy Python ?

In effect np.r_ is doing the same thing as MATLAB [2,4,6:9], expanding and concatenating to create an array of individual indices. MATLAB wraps a lot of functionality in that [...] syntax. In python that syntax is restricted to lists and indexing (which is what r_ is 'pretending' to do!). — hpaulj
– hpaulj, Commented Apr 13, 2021 at 18:36

Mad Physicist · Accepted Answer · 2021-04-13 16:49:14Z

2

You can't combine a fancy integer index and a slice like that in numpy, so you have two options:

Convert the slice to numerical indices:
```
A = np.array([1,2,3,4,5,6,7,8,9])
A[[1, 3, 5, 6, 7, 8]] = -1
```
Notice the double square brackets: the index is a list of integers.
Set the fancy index and slice separately:
```
A[[1, 3]] = A[5:9] = -1
```
Notice that the right bound of a slice is exclusive in python and indices are zero based.

answered Apr 13, 2021 at 16:49

Mad Physicist

116k29 gold badges202 silver badges292 bronze badges

Sign up to request clarification or add additional context in comments.

1 Answer 1