Python RegEx：如何处理行

Question

I have a huge txt file which has the following kind of format:

BadLine
      property1=a
      property2=b
BadLine2
      property1=c
      property2=d
GOODLINE1
      property1=e
      property2=f

....and many more good and bad lines.

What I need to do is to extract the properties of the good lines (e and f in the above example).

I can easily find the good lines in my file, but then how do I select the properties searching other regexs only in the block associated to goodlines?

Thanks guys!

Answer 1

The following code:

import re

test = '''
BadLine
      property1=a
      property2=b
BadLine2
      property1=c
      property2=d
GOODLINE1
      property1=e
      property2=f
BadLine
      property1=a
      property2=b
BadLine2
      property1=c
      property2=d
GOODLINE2
      property1=e
      property2=f
'''

pattern = r'^(GOODLINE(?:[^\n]|\n )*)'

print re.compile(pattern, re.MULTILINE).findall(test)

produces these results:

['GOODLINE1\n      property1=e\n      property2=f', 'GOODLINE2\n      property1=e\n      property2=f']

The pattern matches "GOODLINE" appearing at the beginning of a line, as well as greedily matching characters after it that are not linefeeds, as well as linefeeds that are followed by space characters. If your text actually has tabs after linefeeds instead of spaces, you can change the space into a tab. Alternatively, you could easily match either by changing the pattern like this:

pattern = r'^(GOODLINE(?:[^\n]|\n[ \t])*)'

Once you have these matches, it is extremely easy to use regular string split() in order to extract the properties.

Alternatively, you could see if the rson package parsing satisfies your needs -- this looks like a file it could easily parse.

Answer 2

The short answer is you can us:

GOODLINE[\d+]*\n.*property1=(.+)*\n.*property2=(.+)*\n?

In this case the two parentheses will be the values you are looking for. If you have the string in a file which is created in windows/mac style, you will have different end chars:'\\r\\n' in windows and '\\r' in mac. In linux system you will have '\\n' only. The above pattern will match with any Goodline at the beginning or end of your string, even without any newline at the end. Your values in properties can be more than one character, as well.

You can try a very useful website, Pythex to try your regular expressions.

The code you can try is:

import re
pattern = re.compile('GOODLINE[\d+]*\n.*property1=(.+)*\n.*property2=(.+)*\n?')
matchRes = re.findall(pattern,'''BadLine2
      property1=c
      property2=d
GOODLINE11
      property1=e
      property2=f
BadLine2
      property1=c
      property2=d
GOODLINE11
      property1=eee34
      property2=f00
BadLine2
      property1=c
      property2=d
GOODLINE1
      property1=e
      property2=f''');

if matchRes:
    print matchRes
else:
    print 'No match'

and you will get the following results in a list which each pair is the property1 and property2 values:

[('e', 'f'), ('eee34', 'f00'), ('e', 'f')]