Simple question about numpy matrix in python

Question

Let's suppose I have a numpy matrix variable called MATRIX with 3 coordinates: (x, y, z).

Is acessing the matrix's value through the following code

myVar = MATRIX[0,0,0]

equal to

myVar = MATRIX[0,0][0]

or

myVar = MATRIX[0][0,0]

?

What about if I have the following code?

myTuple = (0,0)
myScalar = 0
myVar = MATRIX[myTuple, myScalar]

Is the last line equivalent to doing

myVar = MATRIX[myTuple[0], myTuple[1], myScalar]

I have done simple tests and it seems so, but maybe that is not so in all the cases. How do square brackets work in python with numpy matrices? Since day one I felt confused as how they work.

Thanks

Are you sure? I get a TypeError when I try any of this. If you have a tuple MYTUPLE=(1,2,3) then the only possible indices are MYTUPLE[0], MYTUPLE[1] and MYTUPLE[2]. — Tim Pietzcker
– Tim Pietzcker, Commented Nov 30, 2009 at 21:16
Sorry, now that I see it, it's numpy matrices that I was refering to. — devoured elysium
– devoured elysium, Commented Nov 30, 2009 at 21:20

Roberto Bonvallet · Accepted Answer · 2009-11-30 21:31:45Z

6

I assume you have a array instance rather than a matrix, since the latter only can have two dimensions.

m[0, 0, 0] gets the element at position (0, 0, 0). m[0, 0] gets a whole subarray (a slice), which is itself a array. You can get the first element of this subarray like this: m[0, 0][0], which is why both syntaxes work (even though m[i, j, k] is preferred because it doesn't have the unnecessary intermediate step).

Take a look at this ipython session:

rbonvall@andy:~$ ipython
Python 2.5.4 (r254:67916, Sep 26 2009, 08:19:36) 
[...]

In [1]: import numpy.random

In [2]: m = numpy.random.random(size=(3, 3, 3))

In [3]: m
Out[3]: 
array([[[ 0.68853531,  0.8815277 ,  0.53613676],
        [ 0.9985735 ,  0.56409085,  0.03887982],
        [ 0.12083102,  0.0301229 ,  0.51331851]],

       [[ 0.73868543,  0.24904349,  0.24035031],
        [ 0.15458694,  0.35570177,  0.22097202],
        [ 0.81639051,  0.55742805,  0.5866573 ]],

       [[ 0.90302482,  0.29878548,  0.90705737],
        [ 0.68582033,  0.1988247 ,  0.9308886 ],
        [ 0.88956484,  0.25112987,  0.69732309]]])

In [4]: m[0, 0]
Out[4]: array([ 0.68853531,  0.8815277 ,  0.53613676])

In [5]: m[0, 0][0]
Out[5]: 0.6885353066709865

It only works like this for numpy arrays. Python built-in tuples and lists are not indexable by tuples, just by integers.

edited Nov 30, 2009 at 21:31

answered Nov 30, 2009 at 21:25

Roberto Bonvallet

33.7k5 gold badges45 silver badges57 bronze badges

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

2 Comments

devoured elysium Over a year ago

hmm, and what would m[0,0,0] yield?

Roberto Bonvallet Over a year ago

The same as m[0, 0][0], given that len(m.shape) == 3.

Alex Gaynor · Accepted Answer · 2009-11-30 21:14:48Z

1

It's not possible to index a tuple with another tuple, so none of that code is valid.

answered Nov 30, 2009 at 21:14

Alex Gaynor

15.1k10 gold badges66 silver badges113 bronze badges

Collectives™ on Stack Overflow

Simple question about numpy matrix in python

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related