def extract_between(content, start_str, end_str):
"""返回 content 中 start_str 與 end_str 之間的子串,不含標記"""
try:
start_idx = content.index(start_str)
# 移動到開始標記之後
start_idx += len(start_str)
end_idx = content.index(end_str, start_idx) # 從 start_idx 開始找結束標記
return content[start_idx:end_idx]
except ValueError:
return None # 標記不存在
# 測試
html = 'jb51.net'
print(extract_between(html, '', '')) # jb51.net
def extract_all(content, start_str, end_str):
results = []
pos = 0
while True:
try:
s = content.index(start_str, pos)
s += len(start_str)
e = content.index(end_str, s)
results.append(content[s:e])
pos = e + len(end_str)
except ValueError:
break
return results
log = '[INFO] start [ERROR] fail [INFO] end'
print(extract_all(log, '[INFO] ', ' ['))
# ['start', 'end']
import re
def extract_re(content, start_pat, end_pat):
pattern = re.compile(re.escape(start_pat) + '(.*?)' + re.escape(end_pat), re.DOTALL)
return pattern.findall(content)
# 同樣效果,支持正則元字符作為邊界
import timeit s = 'content ' * 10000 print(timeit.timeit("extract_between(s, '', ' ')", globals=locals(), number=100)) # ~0.02s print(timeit.timeit("extract_re(s, '', ' ')", globals=locals(), number=100)) # ~0.05s
簡單固定標記用 index() 切片最快;複雜邊界用正則。