viariable changes but no update to it made

Question

I'm trying to flip an image (matrix with RGB values) taken with opencv in Python by switching the columns on the top and bottom:

cv2.imshow('Image',image)
temp = image[0]
row,col,number = image.shape
for i in range (0,col/2) :
        temp = image[i] 
        image[i] = image[col-i-1]
        image[col-i-1] = temp

For some reason, temp changes to image[col-i-1] after executing image[i] = image[dol-i-1].

I printed the first values of each column to see what changed and it showed this:

temp
[ 74 63 167]  (old) 
[ 85 69 177]  (new)
image[i]
[ 85 69 177]  (old)
[ 77 66 170]  (new)
temp
[ 77 66 170]  (old)
[ 77 66 170]  (new)
image[col-i-1]
[ 77 66 170]  (old)
[ 77 66 170]  (new)

So why does temp change?

score 1 · Accepted Answer · edited May 23 '17 at 12:31

The reason the temporary variable seems to change is because in Python, you're not making a deep copy when you assign

temp = image[i]

Instead, you're "binding" a new name to the same object. This means that temp simply points to the same underlying data structure that image[i] is also pointing to. When you next make an update to image[i], that change will also be reflected in the temp variable.

An easy example that shows the same behaviour is this one:

>>> a = [1,2,3]
>>> b = a
>>> a[2] = 0
>>> a  # as expected
[1, 2, 0]
>>> b  # python gotcha
[1, 2, 0]

It's a common pitfall for people who are new to Python, as evidenced from the dozens of related SO questions.

viariable changes but no update to it made

1 Answers1