I am a high school student and looking through some numpy code, I found something along the lines of
a = x[:,0:4]
and x was a 2-d array. I know that a[:], refers to all objects in array a, so for x[:,0:4], would it refer to all rows of x and columns with index 0,1,2,3 excluding column with index 4?
Just trying to get confirmation about how this works because I have seen it in several types of code and just wanted to be sure.
You are right. This a = x[:,0:4] selects the first four columns.
Example:
>>> a = np.arange(25).reshape(5, 5)
>>> a
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]])
You can skip the 0 because a[:,:4] means the same as a[:,0:4]:
>>> a[:,:4]
array([[ 0, 1, 2, 3],
[ 5, 6, 7, 8],
[10, 11, 12, 13],
[15, 16, 17, 18],
[20, 21, 22, 23]])
You can always think: "First dimension first, second dimension second, and so on." In the 2D case the first dimension is the rows and the second dimension is the columns.
Yes this is referred to as slice notation and numpy arrays can also use Python's slice notation, so
>>>x = np.arange(25).reshape(5, 5)
>>>a = x[:, 0:4]
>>>a
array([[ 0, 1, 2, 3],
[ 5, 6, 7, 8],
[10, 11, 12, 13],
[15, 16, 17, 18],
[20, 21, 22, 23]])
If you use the slice notation, x will be a view of a and not a copy, so if you change a value in array x, the value will also be changed in a.
>>>x[1,1] = 1000
>>>a
array([[ 0, 1, 2, 3],
[ 5, 1000, 7, 8],
[ 10, 11, 12, 13],
[ 15, 16, 17, 18],
[ 20, 21, 22, 23]])