X.shape == (10,4)
y.shape == (10)
I'd like to produce M, where each entry in M is defined as M[r,c] == X[r, y[r]]; that is, use y to index into the appropriate column of X.
How can I do this efficiently (without loops)?
M could have a single column, though eventually I need to broadcast it so that it has the same shape as X. c starts from the first col of X (0) and goes to the last (9).
Just do :
X=np.arange(40).reshape(10,4)
Y=np.random.randint(0,4,10)
M=X[range(10),Y]
for
In [8]: X
Out[8]:
array([[ 0, 1, 2, 3],
[ 4, 5, 6, 7],
[ 8, 9, 10, 11],
[12, 13, 14, 15],
[16, 17, 18, 19],
[20, 21, 22, 23],
[24, 25, 26, 27],
[28, 29, 30, 31],
[32, 33, 34, 35],
[36, 37, 38, 39]])
In [9]: Y
Out[9]: array([1, 1, 3, 3, 1, 2, 2, 3, 2, 1])
In [10]: M
Out[10]: array([ 1, 5, 11, 15, 17, 22, 26, 31, 34, 37])
range is a Python object, and, for very large arrays (which I'm using), will be slow. Can we do it using a Numpy array instead?
np.array(range(100000)) is much slower than np.arange(100000), even on python3. Wouldn't be hard to fix that in numpy though
X[np.arange(10),Y] is faster - though it really should be tested on much larger arrays.
M? How is thatcsetup?