Perl Guru Forums: Perl Programming Help: Regular Expressions: Optimising a regex for stripping unwanted whitespace

CGI/Perl Guide | Learning Center | Forums | Advertise | Login

Site Search: in


	MAIN INDEX	SEARCH POSTS	WHO'S ONLINE	LOG IN

Home: Perl Programming Help: Regular Expressions:
Optimising a regex for stripping unwanted whitespace

BorisE
Novice

May 20, 2008, 10:36 AM

Post #1 of 3 (2152 views)

Optimising a regex for stripping unwanted whitespace	Can't Post

and newlines from an XML file?

I have a program that writes data as XML currently using XML::Structured. One side effect is that if I want to write a value <INFLUENCE ID="3">1</INFLUENCE> the tag gets broken across three lines as

Code

     <INFLUENCE ID="3"> 
        1 
    </INFLUENCE>

which is acceptable but undesirable.

I'm currently using the following on the whole text of the XML file and repeating it once for each affected tag type (currently two).

Code

   $rawdata =~ s/(<INFLUENCE\s[^>]+>)\n\s*(\S*)\n\s*(<\/INFLUENCE>)/$1$2$3/g;

Thinking further I know its a long way below optimum and I'm not fond of the use of the "\S*" match in the middle.

Obviously I could use alternation to match all the tag types in one hit but I think I'd be better off testing for any tag with something like

Code

   $rawdata =~ s/(<(\w+)\s[^>]*>)\n\s*([+\-0-9eE]*)\n\s*(<\/\2)/$1$3$4/g;

I'm not happy with the character class [+\-0-9eE] as I want to match any of the characters in a number even if its in scientific notation (I might want to do the same for text) and I'm wondering if I should use a negated class such as [^\n<>] which should accept ANY single line that isn't an XML tag, but I'm worried I could come unstuck with "greed" since it could match the whole line including the whitespace at the beginning if \s* failed to eat it first. I assume that with two greedy matches the first one "wins".

meloyelo
User

Jun 9, 2008, 5:43 AM

Post #2 of 3 (1726 views)

Re: [BorisE] Optimising a regex for stripping unwanted whitespace [In reply to]	Can't Post

Some ideas:

1. use \S*? instead of \S*
2. use two substitute operations:

s{\s*</INFLUENCE>}{</INFLUENCE>}g;
s{(<INFLUENCE.*?>)\s*}{$1}g;

This is less prone to error and leaving you wondering "how come it's not working???"

amodiahs
Novice

Jun 25, 2008, 4:48 AM

Post #3 of 3 (1622 views)

Re: [BorisE] Optimising a regex for stripping unwanted whitespace [In reply to]	Can't Post

Hello,

This is forum related to RegExp, but I have a different soultion to your problem without using RegExp.

You can split your file using "\n" character and hence u will get each line in an array element. Traverse through the array and create a new file with the same XML tags keeping a "\n" whever required...

Search for (options)

Web Applications & Managed Hosting Powered by Gossamer Threads
Visit our Mailing List Archives