In article <mailman.10370.1401191774.18130.python-l...@python.org>,
Wolfgang Maier <wolfgang.ma...@biologie.uni-freiburg.de> wrote:
> On 27.05.2014 13:39, Aman Kashyap wrote:
> >> On 27.05.2014 14:09, Vlastimil Brom wrote:
> >>> you can just escpape the pipe with backlash like any other metacharacter:
> >>> r"start=\|ID=ter54rt543d"
> >>> be sure to use the raw string notation r"...", or you can double all
> >>> backslashes in the string.
> > Thanks for the response.
> > I got the answer finally.
> > This is the regular expression to be
> > used:\\|ID=[a-z]*[0-9]*[a-z]*[0-9]*[a-z]*\\|
> or, and more readable:
> This is what Vlastimil was talking about. It saves you from having to
> escape the backslashes.
Sometimes what I do, instead of using backslashes, I put the problem
character into a character class by itself. It's a matter of personal
opinion which way is easier to read, but it certainly eliminates all the
questions about "how many backslashes do I need?"
Another thing that can help make regexes easier to read is the VERBOSE
flag. Basically, it ignores whitespace inside the regex (see
https://docs.python.org/2/library/re.html#module-contents for details).
So, you can write something like:
pattern = re.compile(r'''[|]
Or, alternatively, take advantage of the fact that Python concatenates
adjacent string literals, and write it like this:
pattern = re.compile(r'[|]'