I have an 8000-element 1D array.
I want to obtain the following two arrays:
test contains the element with the index from [1995:1999], [3995:3999], [5999:5999], [7995:7999].
train should contains everything else.
How should I do that?
idx = [1995,1996,1997,1998, 1999, 3995, 3996, 3997,3998, 3999, 5995, 5996, 5997, 5998, 5999, 7995, 7996, 7997, 7998, 7999]
test = [X[i] for i in idx]
train = [X[i] for i **not** in idx]
Based on your example, a simple workaround would be this:
train = [X[i] for i, _ in enumerate(X) if i not in idx]
I looks like you are looking for numpy.where, here is a simple example to get you started:
In [18]: import numpy as np
In [19]: a = np.array([[0,3],[1,2],[2,3],[3,2],[4,5],[5,1]])
In [20]: a[np.where((a[:, 0] > 1) & (a[:, 0] < 5))[0]]
Out[20]:
array([[2, 3],
[3, 2],
[4, 5]])
In [21]: a[np.where(~((a[:, 0] > 1) & (a[:, 0] < 5)))[0]]
Out[21]:
array([[0, 3],
[1, 2],
[5, 1]])
The first element in row can be your index, and second your value. numpy.where checks whether condition is true or false, and returns a binary array (actually tuple of arrays), once we have binary array, we can index the original array based on that.
If you want, you can use masks
mask = np.ones(len(X), dtype=bool)
mask[idx] = False
train = X[mask]
test = X[idx]
# you can also use this for test
test = X[np.logical_not(mask)]
mask. You negate the mask which is bool. idx is integer type.When building train, you need to iterate through all of your source data.
Using enumerate should make things easy:
>>> data = list(range(8000))
>>> train, test = [], []
>>> for i, value in enumerate(data):
... if 1995 <= i <= 1999 or 3995 <= i <= 3999 or 5995 <= i <= 5999 or 7995 <= i <= 7999:
... test.append(value)
... else:
... train.append(value)
...
>>> test
[1995, 1996, 1997, 1998, 1999, 3995, 3996, 3997, 3998, 3999, 5995, 5996, 5997, 5998, 5999, 7995, 7996, 7997, 7998, 7999]
>>> len(train)
7980
This is one possibility, assuming array is the name of the list containing 8000 elements:
idx = {1995, 1996, 1997, 1998, 1999, 3995, 3996, 3997, 3998, 3999, 5995, 5996, 5997, 5998, 5999, 7995, 7996, 7997, 7998, 7999}
test = [array[x] for x in idx]
train = [x for i, x in enumerate(array) if i not in idx]
vstackfrom numpy. What have you tried so far?