Capturing string in Python

Question

So, I'm trying to capture this big string in Python but it is failing me. The regex I wrote works fine in regexr: http://regexr.com/3cmdc

But trying to using it in Python to capture the text returns None. This is the code:

pattern = "var initialData = (.*?);\\n"
match = re.search(pattern, source).group(1)

What am I missing ?

Are you sure the input has no linebreaks as in the regex demo? If you have linebreaks, you can try match = re.search(pattern, source, re.S).group(1), but if your string is very large, you might have an issue related to the lazy matching stack "overflow". — Wiktor Stribiżew, Commented Jan 28, 2016 at 23:32
Try r'var initialData = ([^;]*(?:;(?!\\n)[^;]*)*)' pattern if your \n contains a literal \. Else, if \n is a normal linebreak, I'd advise r'var initialData = ([^;]*(?:;(?!\n)[^;]*)*)' — Wiktor Stribiżew, Commented Jan 28, 2016 at 23:37
Consider paring down your example string. The problem is easily demonstrated with a much smaller example that could be added to the question directly. — tdelaney, Commented Jan 29, 2016 at 0:06
There is only one line break in that sample string. Just why do you care about it though ? var initialData = (.*?); or var initialData = (.*?);\r?\n and get on with life. Or even var initialData = ([\S\s]*?); — user557597, Commented Jan 29, 2016 at 0:16

alecxe · Accepted Answer · 2016-01-28 23:28:19Z

1

You need to set the appropriate flags:

re.search(pattern, source, re.MULTILINE | re.DOTALL).group(1)

answered Jan 28, 2016 at 23:28

alecxe

475k127 gold badges1.1k silver badges1.2k bronze badges

1

Why use re.MULTILINE | re.DOTALL? OP does not use them at regexr but gets a match. And there are no ^/$ here, so re.M is totally redundant.
– Wiktor Stribiżew
Commented Jan 28, 2016 at 23:30
@WiktorStribiżew well, I suspect the newlines on regexr are not handled properly. Adding flags actually worked for me. Let's see if the answer is gonna help the OP. If not, I'll definitely remove the answer. Thanks.
– alecxe
Commented Jan 28, 2016 at 23:50
Most probably the \n in the regex demo are real linebreaks, so re.S is OK to use, I believe. Although OP says the strings are huge, and I think lazy dot matching is very fragile here.
– Wiktor Stribiżew
Commented Jan 28, 2016 at 23:51

Add a comment |

Jannik Becher · Accepted Answer · 2016-01-28 23:44:38Z

1

Use pythons raw string notation:

pattern = r"var initialData = (.*?);\\n"
match = re.search(pattern, source).group(1)

answered Jan 28, 2016 at 23:44

Jannik Becher

1891 gold badge1 silver badge10 bronze badges

1

It would be helpful to mention why... to match a literal \n in the text, the regex needs to escape the backslash to \\n but then python either needs a raw string r"\\n" or extra escaping of the backslashes "\\\\n".
– tdelaney
Commented Jan 29, 2016 at 0:04

Add a comment |

2 Answers 2