如何按条件将列表拆分为列表？

Question

I want split a list into multiple lists, based on a condition. If the difference in the series grows more than 4, then it is trigger to split the list till that item from last split item.

For example:

in = [1,2,3,9,10,11,100,200]
out = [ [1,2,3 ], [ 9,10,11 ], [100], [200]  ]

by condition

If (next - prev) > 4

Answer 1

def splitlist(L):
    if not L: return []
    answer = [[L[0]]]

    for i in L[1:]:
        if i - answer[-1][-1] < 4:
            answer[-1].append(i)
        else:
            answer.append([i])
    return answer

Output:

In [112]: splitlist([1,2,3,9,10,11,100,200])
Out[112]: [[1, 2, 3], [9, 10, 11], [100], [200]]

Answer 2

NOTICE: in and next are Python keywords and builtin functions respectively.

The trick is to use zip() function with 2 slices of the same list, but shifted on one item:

in_list = [1, 2, 3, 9, 10, 11, 100, 200]

parts = [[]]
for prev_item, next_item in zip(in_list[:-1], in_list[1:]):
    top = parts[-1]
    top.append(prev_item)
    if next_item - prev_item > 4:
        parts.append([])
top = parts[-1]
top.append(in_list[-1])

The output is:

[[1, 2, 3], [9, 10, 11], [100], [200]]

Answer 3

Short solution using numpy module:

import numpy as np

arr = np.array([1,2,3,9,10,11,100,200])
out = [a.tolist() for a in np.split(arr, np.where(np.diff(arr) > 4)[0]+1)]
print(out)

The output:

[[1, 2, 3], [9, 10, 11], [100], [200]]

---

• np.where(np.diff(arr) > 4) - find the array indices where condition "difference between next value and previous value is greater than 4" is met

• np.split(x, indices) - split the initial array by crucial indices

Answer 4

One of the possible approaches:

ls = [1,2,3,9,10,11,100,200, 201, 209]

#output list
output = list()
startind = 0

for ind, val in enumerate(ls):
    if ls[ind]-ls[ind-1] > 4:
        output.append(ls[startind:ind])
        startind = ind

# appending the left over from last detected difference of 4
output.append(ls[startind:])
print output

Output: [[1, 2, 3], [9, 10, 11], [100], [200, 201], [209]]