您的實現可能會爲它正在嘗試解決的問題嵌套太多的循環。
該第一個實現包含錯誤。請參閱下面的修復。
嘗試一些沿也許這些線路:
l = [97,122,111,98,111,98,101,103,103,104,97]
out = []
acc = []
for v in l:
if len(acc)==0 or v >= acc[-1]:
acc.append(v)
else:
if len(acc) > 1:
out.append(acc)
acc = [v]
print(out)
>>>[[97, 122], [98, 111], [98, 101, 103, 103, 104]]
以前的代碼是緩慢的,可以去掉最後發現片段。我發現運行隨機測試時出現錯誤,嘗試優化版本。下面的代碼顯示了原始代碼和更正版本,優化版本可以快30%。
def original(l):
out = []
acc = []
added = False
for v in l:
if len(acc)==0 or v >= acc[-1]:
acc.append(v)
else:
added = False
acc = [v]
if acc is not None and len(acc)>1 and not added:
added = True
out.append(acc)
return out
def optimized(l):
out = []
acc = None
tmp = None
deb_v = False
for v in l:
prev = acc[-1] if (acc is not None and len(acc)) else tmp
if prev is not None and v >= prev:
if tmp is not None:
acc = []
acc.append(tmp)
out.append(acc)
tmp = None
acc.append(v)
else:
acc = None
tmp = v
return out
# The original test data
l = [97,122,111,98,111,98,101,103,103,104,97]
assert original(l) == optimized(l) == [[97,122],[98,111],[98,101,103,103,104]]
# A list that triggered last-fragment-dropped error
l = [57, 16, 6, 19, 40, 3, 4, 13, 2, 70, 85, 65, 32, 69, 54, 51, 95, 74, 92, 46, 45, 26, 0, 61, 99, 43, 67, 71, 97, 10, 18, 73, 88, 47, 33, 82, 25, 75, 93, 80, 23, 37, 87, 90, 49, 15, 35, 63, 17, 64, 5, 72, 89, 21, 50, 8, 41, 86, 31, 78, 52, 76, 56, 42, 77, 36, 11, 60, 39, 22, 68, 27, 24, 28, 59, 96, 29, 38, 12, 79, 53, 9, 83, 94, 34, 14, 7, 48, 30, 20, 66, 62, 91, 58, 81, 1, 98, 44, 55, 84]
assert original(l) == optimized(l)
# Random testing
import random
l = list(range(100))
random.shuffle(l)
assert original(l) == optimized(l)
# Timing!
import timeit
print(timeit.timeit("original(l)", globals={"l":l, "original": original}))
# 43.95869998800117
print(timeit.timeit("optimized(l)", globals={"l":l, "optimized": optimized}))
# 34.82134292599949
您解釋了'for'循環的行爲。但你試圖解決的問題陳述並不清楚。請編輯問題以解釋所需的行爲。另外提及樣本輸入和期望的輸出 –
您的代碼中除了您聲明的代碼之外的幾個問題。 'a [m]'和'a [n]'是字符串,你在哪裏首先定義'c'? – Tagc
你想做什麼?迭代時修改列表幾乎總是一個壞主意 –