I have a big np.ndarray (3600000,3), the HUE, the VALUE, and an associated CLASS number. For each pairs of HUE and VALUE I would like to find, using this array the corresponding Class number. I'm a very beginner in Python and have a hard time doing it. Do you know a way to do it?
Thank you in advance!
I assume your array looks like:
|(HUE)(VALUE)(CLASS)
row/col| 0 1 2
-------+-----------------
0 | 0 1 2
1 | 3 4 5
2 | 6 7 8
. | . . .
. | . . .
3599999| . . .
And here is the sample code. For simplicity I changed the size 3600000 to 5.
a = np.array(xrange(5 * 3))
a.shape = (5, 3)
Now array a look like this:
array([[ 0, 1, 2],
[ 3, 4, 5],
[ 6, 7, 8],
[ 9, 10, 11],
[12, 13, 14]])
If you want row with HUE=9, do like this:
a[np.where(a[:,0] == 9)]
#array([[ 9, 10, 11]])
If you want row with VALUE=4, do like this:
a[np.where(a[:,1] == 4)]
#array([[3, 4, 5]])
If you want row with HUE=0 and VALUE=1, do like this:
a[np.where((a[:,0] == 0) * (a[:,1] == 1))]
#array([[0, 1, 2]])
HUE=1 or VALUE=1 ?Try this code:
x[x[:, 2] == class_number[:, :2]
where x is np.ndarray
x[:, 2] == class_number
contains true/false that means whether the last is class_number or not.
You need to take a look at: Boolean indexing in http://wiki.scipy.org/Cookbook/Indexing
Moved from comment.