Fast methods to convert list to set? python 2.7

Question

so in my current code I'm using

im_set=set(map(tuple, im_list[0]))

to convert a list into a set. However it's very slow. Is there a faster/ none map method to speed things up? This is my current code

while True:
    cam = cv2.VideoCapture(0)
    start=time.time()
    while(cam.isOpened()):                  #Opens camera
        ret, im = cam.read()                #Takes screenshot
    #im=cv2.imread('RGB.png')
        im=cv2.resize(im,(325,240))         #Resize to make it faster
        im= im.reshape(1,-1,3)
    #im=OneNumber(im)               #Converts the pixels rgb to a singe number
        im_list=im.tolist() 
        im_set=set(map(tuple, im_list[0]))
        ColourCount= set(im_set) & set(om_set)
        print len(ColourCount)

    print N
    end=time.time()-start
    print end

Note that om_set is from a different program.

Anyway, basically I have to convert an np.array to a set, and compare that set to see what overlaps with them. But, its very slow. Is there a method I can use to speed up this conversion?

score 0 · Answer 1 · edited May 11 '16 at 11:31

0

If you were using Python 3, that would be fast; but in Python 2, you should use a set comprehension:

im_set = {tuple(item) for item in im_list[0]}

map() is slower in Python 2 because it returns a list. That means that each tuple is stored in memory before you even start creating a set. This way, however, you are adding each item to the set as it is converted to a tuple.

edited May 11 '16 at 11:31

Padraic Cunningham

176,452
29
245
321

answered May 10 '16 at 10:42

zondo

19,901
8
44
83

Am I able to get cv2, bumpy and PIL in python 3? – Davey Boy May 10 '16 at 13:53
I believe you can get cv2 and PIL. I don't know about bumpy, but I would guess so. The set comprehension is just as fast, and maybe faster, than `map()`, so the switch to Python 3 isn't urgent. – zondo May 10 '16 at 14:03
1

It was mine accidentally – Padraic Cunningham May 11 '16 at 11:33

score 0 · Answer 2 · edited May 23 '17 at 10:28

0

Using itertools.imap is going to be faster:

In [18]: from itertools import imap

In [19]: timeit set(imap(tuple, l))
100 loops, best of 3: 2.27 ms per loop

In [20]: timeit {tuple(item) for item in l}
100 loops, best of 3: 2.69 ms per loop

You also don't need to create extra sets, you use the ones you have created and use set.intersection:

im_set.intersection(om_set)

Unless you specifically need lists this can also all be done using numpy.

edited May 23 '17 at 10:28

Community

1
1

answered May 11 '16 at 11:31

Padraic Cunningham

176,452
29
245
321

This is probably a dumb question, but what does In [x]: mean? Also for the array intersecting, will this show if any pixel types match up or only if a row of pixels matches up? – Davey Boy May 13 '16 at 15:01
@DaveyBoy, that is just output from my ipython shell, `inter = set(imap(tuple, im_list[0])).intersection(om_set)` will match your own logc but if om_set is actually a set just use `om_set.intersection(imap(tuple, im_list[0]))`. – Padraic Cunningham May 13 '16 at 15:25

Fast methods to convert list to set? python 2.7

2 Answers2