如何在正则表达式中匹配'或'与'或'

Question

The following regular expression is used to extract the URL link from a page:

LINK_REGEX = re.compile("<a [^>]*href=['\"]([^'\"]+)['\"][^>]*>")

Question1 > How to represent the following string? I mismatch the ' and " in purpose

<a href="http://www.yahoo.com'>

I have tried the following statements and none work for me.

>>> page = '<a href="http://www.yahoo.com\'>'
>>> page
'<a href="http://www.yahoo.com\'>'
>>> page = '<a href="http://www.yahoo.com''>'
>>> page
'<a href="http://www.yahoo.com>'

Question2 > Based on my understanding, by design, the LINK_REGEX will match above link although this is not desirable. So how can I modify the regular expression so that it enforces the matching of ' with ' or " with ".

Answer 1

For Question 1, your first approach works.

>>> page = '<a href="http://www.yahoo.com\'>'
>>> len(page)
31
>>> page
'<a href="http://www.yahoo.com\'>'
>>> page[-1]
'>'
>>> page[-2]
"'"
>>> page[-3]
'm'

Answer 2

(I'd post this as a comment if I had the privilege.)

If you're trying to parse HTML, it is highly recommended that you do not use regex. You'll be saving yourself lots of hassle and problems if you use an HTML parsing module like BeautifulSoup or lxml.html.

Second, pretty much every time you're using regex, be sure to prepend r to your string, like so:

LINK_REGEX = re.compile(r"<a [^>]*href=['\\"]([^'\\"]+)['\\"][^>]*>")

This will ensure things are escaped properly.

If you definitely need to use regex though, "9000's" answer will work for you.

Answer 3

['"] will match ' or " .

(['"]).+\\1 will match a quoted string with matcing quotes. The expression in parens (match group) will match a single or double quote, and \\1 will match whatever the first match group have matched (this is called 'backreference').

Note that the quotes are not escaped in any way in the expressions to make them more readable. Your regex strings may need to escape at least one kind of quotes.

Answer 4

Use two regexes:

<a\s*[^>]*href="([^"]+)"[^>]*>  # double quoted strings
<a\s*[^>]*href='([^']+)'[^>]*>  # single quoted strings

The content of href will then be in the second group.