splitting by double newline

[Nikola Skoric]

Hello everybody,

I'd like to split a file by double newlines, but portably. Now,
splitting by one or more newlines is relatively easy:

self.tables = re.split("[\r\n]+", bulk)

But, how can I split on double newlines? I tried several approaches,
but none worked...

 

[Ian]

Hello everybody,

I'd like to split a file by double newlines, but portably. Now,
splitting by one or more newlines is relatively easy:

self.tables = re.split("[\r\n]+", bulk)

But, how can I split on double newlines? I tried several approaches,
but none worked...

self.tables = re.split(r'(?:\r\n){2,}|\r{2,}|\n{2,}', bulk)

 

[Peter Otten]

Hello everybody,

I'd like to split a file by double newlines, but portably. Now,
splitting by one or more newlines is relatively easy:

self.tables = re.split("[\r\n]+", bulk)

But, how can I split on double newlines? I tried several approaches,
but none worked...

If you open the file in universal newline mode with

with open(filename, "U") as f:
bulk = f.read()

your data will only contain "\n". You can then split with

blocks = bulk.split("\n\n") # exactly one empty line

or

blocks = re.compile(r"\n{2,}").split(bulk) # one or more empty lines

One last variant that doesn't read in the whole file and accepts lines with
only whitespace as empty:

with open(filename, "U") as f:
blocks = ("".join(group) for empty, group in itertools.groupby(f,
key=str.isspace) if not empty)

 

[Nikola Skoric]

Dana Mon, 07 Feb 2011 19:20:38 +0100,

with open(filename, "U") as f:

Oh, fabulous. Thanks!

 

[Nikola Skoric]

Dana Mon, 7 Feb 2011 10:02:05 -0800 (PST),

self.tables = re.split(r'(?:\r\n){2,}|\r{2,}|\n{2,}', bulk)

Thanks!

I tried without "?:", but it didn't work. Can you tell me why is it
important that group is noncapturing?

 

[MRAB]

Dana Mon, 7 Feb 2011 10:02:05 -0800 (PST),


Thanks!

I tried without "?:", but it didn't work. Can you tell me why is it
important that group is noncapturing?

The scanner uses capture groups to identify which part matched. If you
provide additional capture groups it has problems...