Matching blank lines with regular expressions

Question

I've got a string that I'm trying to split into chunks based on blank lines.

Given a string s, I thought I could do this:

re.split('(?m)^\s*$', s)

This works in some cases:

>>> s = 'foo\nbar\n \nbaz'
>>> re.split('(?m)^\s*$', s)
['foo\nbar\n', '\nbaz']

But it doesn't work if the line is completely empty:

>>> s = 'foo\nbar\n\nbaz'
>>> re.split('(?m)^\s*$', s)
['foo\nbar\n\nbaz']

What am I doing wrong?

[python 2.5; no difference if I compile '^\s*$' with re.MULTILINE and use the compiled expression instead]

Answer 1

Looks like this functions as designed. From http://docs.python.org/library/re.html

Note that split will never split a string on an empty pattern match. For example:

>>> re.split('x*', 'foo')
['foo']
>>> re.split("(?m)^$", "foo\n\nbar\n")
['foo\n\nbar\n']

You may want to try building something around re.finditer instead.

Answer 2

What you're doing wrong is using regular expressions. What is wrong with ('Some\ntext.').split('\n')?

Answer 3

Try this instead:

re.split('\n\s*\n', s)

The problem is that "$ *^" actually only matches "spaces (if any) that are alone on a line"--not the newlines themselves. This leaves the delimiter empty when there's nothing on the line, which doesn't make sense.

This version also gets rid of the delimiting newlines themselves, which is probably what you want. Otherwise, you'll have the newlines stuck to the beginning and end of each split part.

Treating multiple consecutive blank lines as defining an empty block ("abc\n\n\ndef" -> ["abc", "", "def"]) is trickier...

Answer 4

Is this what you want?

>>> s = 'foo\nbar\n\nbaz'
>>> re.split('\n\s*\n',s)
['foo\nbar', 'baz']

>>> s = 'foo\nbar\n \nbaz'
>>> re.split('\n\s*\n',s)
['foo\nbar', 'baz']

>>> s = 'foo\nbar\n\t\nbaz'
>>> re.split('\n\s*\n',s)
['foo\nbar', 'baz']