问题
I have a list of lists like this:
L=[[[1,2,3],[4,5]],[[6,7,8,9],[10]]]
I want to append the integer 11 to the subsublists 1 and 3. I can do something like:
L[0][2].append(11)
L[1][2].append(11)
Is there a simpler way to do it in Python ?
Because in my case, let's say I have a list with 100 sublists, and these sublists have 100 sublists (comparable to a (100,100)-matrix) and I want to append a value to the sublists from nb 50 to 75 of the sublists from nb 10 to 20.
So right now I do something like:
for i in range(10,21):
for j in range(50,76):
L[i][j].append(value)
Is there a more efficient way ? Like with numpy arrays we can do
L=[10..21,50..76]=value
回答1:
how to use numpy arrays in this case since L[i][j].size changes with i and j. Is it possible to use arrays in this case ?
Yes, but the dtype is object in such case.
L=[[[1,2,3],[4,5]],[[6,7,8,9],[10]]]
L=np.array(L) # L is a ndarray of list
# array([[[1, 2, 3], [4, 5]], [[6, 7, 8, 9], [10]]], dtype=object)
value=100
for i in L[0:1,0:2].flatten():
i.append(value)
# array([[[1, 2, 3, 100], [4, 5, 100]], [[6, 7, 8, 9], [10]]], dtype=object)
In this example, L is a numpy.ndarray of python list objects.
type(L)
# <type 'numpy.ndarray'>
type(L[0,0])
# <type 'list'>
Arithmetic operation on jagged array
It is possible to perform efficient arithmetic operation on the jagged array like L using numpy.
marr = np.vectorize(np.array,otypes=[np.ndarray])
L=[[[1,2,3],[4,5]],[[6,7,8,9],[10]]]
L=marr(L) # L is a ndarray of ndarray
L+L
# array([[array([2, 4, 6]), array([ 8, 10])],[array([12, 14, 16, 18]), array([20])]], dtype=object)
来源：https://stackoverflow.com/questions/36981925/append-the-same-element-to-several-sublists-in-python