Filtering with itemgetter and numpy arrays

Question

I'm getting the following error:

Traceback (most recent call last):
  File "calibrating.py", line 160, in <module>
    intrinsic = calibrate2(corners, cb_points, (640,480))
  File "calibrating.py", line 100, in calibrate2
    valid_corners = filter(itemgetter(0), image_corners)
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

image_corners is a list of numpy arrays, i.e.,

[array([[ 261.45239258,  140.88212585],
   [ 301.11242676,  156.306427  ],
   [ 343.38937378,  168.20132446],
   [ 382.79559326,  180.48405457],...
   [ 392.16989136,  338.6171875 ],
   [ 439.97772217,  337.2124939 ]], dtype=float32), ... ]

What I want to do is to take the matrices without the dtype=float32, what am I doing wrong?

Could you clarify what you mean by "take the matrices without the dtype=float32"? — Janne Karila, Apr 29 '13 at 09:13
What is a "valid corner"? Is it any point other than `[0,0]`? (What are you trying to eliminate with `filter`?) — unutbu, Apr 29 '13 at 09:21
With "take the matrices without the dtype=float32" I mean that I want to take out the second element of the numpy array tupple, in other words, get a "traditional" array. — XavierusWolf, Apr 29 '13 at 09:45
Valid corners is the name I gave to the list which contains matrices that I can operate with. — XavierusWolf, Apr 29 '13 at 09:46
What is a "traditional" array? Do you mean you want to convert all the NumPy arrays to eg. lists of lists of float? — Janne Karila, Apr 29 '13 at 09:49
See [Convert 2d numpy array into list of lists](http://stackoverflow.com/q/9721884/222914) — Janne Karila, Apr 29 '13 at 09:51

score 1 · Answer 1 · answered Apr 29 '13 at 09:19

1

The dtype attribute isn't accessible by itemgetter.

Try this filter instead:

filter(lambda arr: arr.dtype != float32, image_corners)

That will give you all the matricies without dtype==float32.

answered Apr 29 '13 at 09:19

Raymond Hettinger

216,523
63
388
485

score 0 · Answer 2 · answered Apr 29 '13 at 09:26

0

Or with list comprehensions:

[a for a in image_corners if a.dtype is not float32]

answered Apr 29 '13 at 09:26

Jakub M.

32,471
48
110
179

Filtering with itemgetter and numpy arrays

2 Answers2