如果下一行与模式不匹配，则连接 Python 中的行

Question

您好，我有如下所示的加演字幕文件：

00:00:29:02 00:00:35:00 text 1
text 2
00:00:36:04 00:00:44:08 text 3
text 4
00:00:44:12 00:00:48:00 text 5
00:00:49:17 00:00:52:17 text 6

在 python 我应该放什么而不是 "HELP PLEASE"

newdata = re.sub("""HELP PLEASE""", r"", filedata)

生成这样的行：

00:00:29:02 00:00:35:00 text 1 text 2
00:00:36:04 00:00:44:08 text 3 text 4
00:00:44:12 00:00:48:00 text 5
00:00:49:17 00:00:52:17 text 6

谢谢

Answer 1

如果文件不是太大，您可以将每一行读入一个新列表。如果一行不以时间戳开头，则弹出添加到 new_lines 的最后一行并将其添加回来并附加新行。

>>> import re
>>>
>>> # assume all_lines = somefile.readlines() or use it in the for loop below.
... # but simplying to this
... all_lines = [
... "00:00:29:02 00:00:35:00 text 1",
... "text 2",
... "00:00:36:04 00:00:44:08 text 3",
... "text 4",
... "00:00:44:12 00:00:48:00 text 5",
... "00:00:49:17 00:00:52:17 text 6",
... "text 7",  # added for interest
... "text 8",  # added for interest
... ]
>>>
>>> new_lines = []
>>> for line in all_lines:
...     if not re.match('(?:(?:\d\d:){3}(?:\d\d) ){2}.*', line):
...         # line did not start with a timestamp
...         new_lines.append(new_lines.pop() + ' ' + line)
...     else:
...         new_lines.append(line)
...
>>> print '\n'.join(new_lines)
00:00:29:02 00:00:35:00 text 1 text 2
00:00:36:04 00:00:44:08 text 3 text 4
00:00:44:12 00:00:48:00 text 5
00:00:49:17 00:00:52:17 text 6 text 7 text 8
>>>

使用您保留 dumping/yielding 的 prev_line 变量而不是潜在的巨大 new_lines.

变量应该不会太难

顺便说一句，如果第一行不是时间戳，这将失败。

PS：不知道为什么每个人都对正则表达式如此感兴趣。

编辑：无需创建可能庞大的 new_lines 列表...

>>> prev_line = ''
>>> for line in all_lines:
...     if not re.match('(?:(?:\d\d:){3}(?:\d\d) ){2}.*', line):
...         prev_line += ' ' + line
...     else:
...         if prev_line:  # prevents the first flag '' prev_line from printing
...             print prev_line
...         prev_line = line
...
00:00:29:02 00:00:35:00 text 1 text 2
00:00:36:04 00:00:44:08 text 3 text 4
00:00:44:12 00:00:48:00 text 5
>>> print prev_line  # make sure to print/dump the last one
00:00:49:17 00:00:52:17 text 6 text 7 text 8
>>>

两个警告：(1.) 如果一行实际上是空白，它将被跳过。 (2.) 虽然带有 prev_line 的第二个版本即使文件很大也是内存高效的，但是如果你有 many 没有时间戳的连续行它会占用内存（比如第 7 行和第 8 行）- prev_line 必须保留所有内容，直到出现带有时间戳的行。您可以通过转储到一个文件来解决它，没有明确的换行符（\n）并在转储以时间戳开头的行之前添加一个换行符。

如果下一行与模式不匹配，则连接 Python 中的行

concatenate lines in Python if next line not matches pattern

python

regex

negative-lookahead