How do I extract the text before and after a known marker with a Perl regex?

Question

Can anybody tell me how to identify the middle part interestedInThis and backreference the prefix: fontsize=12 and postfix: fontstyle=bold as ${1} and ${2}?

I'm dealing with this string:

<fontsize=12 interestedInThis fontstyle=bold>

Addendum: Sorry, I was not precise enough, here are the specifics:

• prefix and postfix could be absent
• prefix and postfix can be any string, not necessarily fontsize, resp. fontstyle
• I know for sure, what I am looking for, namely interestedInThis and it will be separated through whitespaces.

Answer 1

<([^>]*)interestedInThis([^>]*)>

Answer 2

For your example, this could work

(<fontsize=\d+) (\w+) (fontstyle=bold>)

Unfortunatly, Perl doesn't seem to support named backreferences so I think you are stuck with <fontsize=12 in $1, ImInterestedInThis in $2 & fontstyle=bold> in $3.

regards, Lieven

Answer 3

Basically

(<fontsize=12) (\S*) (fontstyle=bold>)

But, will the attribute values change? And, do you have to account for variable whitespace? If so, the above mutates into:

(<fontsize=\d+)\s+(\S*)\s+(fontstyle=.*>)

Also, in the above, by using \S, interestedInThis can contain anything that is not whitespace. If there is whitespace there too, for example interestedInThis is actually something like class="x" id="y", then maybe:

(<fontsize=\d+)(.*)(fontstyle=.*>)

Note that $2 is interestedInThis, and $1/$3 is actually your end pieces.

Answer 4

Try this:

my $result = m/(.*)(InterestedInThis)(.*)/;

Now:

• $result is true if it found a match to the format.
• InterestedInThis is in $2, though you already know what it is.
• prefix (EVERYTHING before "InterestedInThis") is in $1.
• postfix (EVERYTHING after "InterestedInThis") is in $3.

Answer 5

I think this is what you want;

<(.* )?InterestedInThis( .*)?>

It will return the the pre- and post-fix if they're there, but will still match if only one or neither are present.

It does have a minor problem that the spaces will be included in the tagged expressions, but that should be easy to remove after the match.

Alternatively, you could use lookahead / lookbehind to try to filter the spaces out as part of the match:

<(.*(?= ))? ?InterestedInThis ?((?<= ).*)?>