occurances of a word in a DCR

So, the division of the company where I work will be sold. We have about 20K DCRs that need to be changed. Since I am nervous about doing that automatically (I will probably have to anyway) I am trying to get the scope of the problem

So what I am trying to do is slurp in a DCR, look for case in sensitive Company and get 3 words on either side.

I am stuck on a regex to get that info. Can someone point me
in the right direction ?

Andy

Find more posts tagged with

Comments

JonathonG

Is this what you're looking for?

$str = 'Here are the words before Company and the words after it';
$str =~ /\W(\w+\W\w+\W\w+\Wcompany\W\w+\W\w+\W\w+\W)/i;
print "Match is: $1\n";

Output is:
Match is: the words before Company and the words

Jonathon
Independent Interwoven Contractor

nipper

Thanks Jonathon, close, one thing I need ot look at is that some of the DCRs were created before the "print pretty DCR" flag was set. So they are all in one line.

That being said, I will need to hack this to work for multiple instances of the pattern.

Andy

JonathonG

What do you want to do with these words after you find them? To return them, you could just use:

@matches = ( $str=~ /INSERTREGEXHERE/g );

Then, each match is put in an entry in the list.

More information here:
http://www.perldoc.com/perl5.6.1/pod/perlretut.html#Using-regular-expressions-in-Perl
(Scroll down to the section discussing //g and \G)

This also may help:

http://www.perldoc.com/perl5.6.1/pod/perlfaq6.html#What-good-is-%5cG-in-a-regular-expression-

Jonathon
Independent Interwoven Contractor

Johnny

If you are concerned about the physical layout of the xml structure, I would certainly use a parser, something like TeamSite: Smiley Very Happy

CRnode .... run through the document and apply the regex to the content (PCDATA)

John Cuiuli

Adam Stoller

I'm a bit confused as to the "before" and "after" desired here.

The suggestions thus far seem to be along the lines of take:
"A B C D Company E F G H" and return "B C D Company E F G"

Is this what you're looking for?

My first impression on reading this was that you were looking to *insert* 3 words on either side of "Company" - i.e. =~ s/Company/X Y Z Company Z Y X/g;

Alternatively, you might be wanting to change the three words surrounding Company - i.e.: =~ s/(\W)\w+\W+\w+\W+\w+\W+Company\W+\w+\W+\w+\W+\w+(\W+)/$1 New Name $2/gs;
(assumes that you've slurped everything into a single string and that I got my w's and W's right)

The use of a parser (TeamSite::XMLnode, etc.) helps to restrict where you're looking to perform whatever regex you are trying to do - since I don't think you'd want to munge something like <tag attr="X">A B Company C D</tag> just in case it's possible to have less than 3 words preceding or following "Company".

If you can provide a more concise description of before/after examples it would probably help us help you (unless we already have).

--fish
Senior Consultant, Quotient Inc.
http://www.quotient-inc.com

nipper

Sure fish,

Basically I will be asked to use a perl script to convert Company to NewCompany, however I am fairly certain that this will not be sufficient.

So right now I am not looking to change anything, just produce a report that says:

DCR1
http://www.company.com/
mailto:andy@company.com
Company is proud to announce......
This division of Company has released ....
(Security Level) Company General Business Information

There are about 18K DCRs in with 9 different categories I need to search. Also, not all DCRs are "pretty" so the whole
DCR is on one line.

This can give me an idea how difficult changing COmpany name will be.

Really I am looking for some word with a little context (which is why I used an arbitrary 3 words before 3 after).

TIA

Andy

Capture.JPG

Adam Stoller

It's not elegant, but I believe the following will do what you want:

sub check_file {
    my($file, $company) = @_;
    if (!open(IN, "<$file")){
        warn(" '$file' ($!)");
        return;
    }
    my @text = ;
    my $count = 0;
    my @found = ();
    foreach my $line (@text){
        chomp $line;
        next if ($line =~ /^\s*$/);
        my $match = ($line =~ m|\b$company\b|i);
        if ($match){
            my @parts = split(/\s+/, $line);
            my $max = scalar(@parts);
            my($i, $s, $e) = (0,0,0);
            foreach(@parts){
                $s++ if ($i > 3);
                $e = $i + 3;
                $e-- while($e > $max);
                if ($parts[$i] =~ /(\b$company\b)/i){
                    $count++;
                    push(@found, join(" ", @parts[$s..$e]));
                }
                $i++;
            }
        }
    }
    print "$file: $count\n\t", join("\n\t", @found), "\n" if ($count);
}

I tried various regex's and couldn't get one that worked on all the cases you provided (and a few of my own) - so I opted for a more brute-force method.

--fish
Senior Consultant, Quotient Inc.
http://www.quotient-inc.com

Migrateduser

Here's an awesome and easy way of doing it. We use it all the time.

Use the program SR32.exe (Search and Replace 32). It works great. It will give you a list of the files it found and what line will be changed. You can then have it overwrite those files or save them into another location. You can also out put the list into a file so that you can go back into Teamsite and submit them through your normal workflows.

Get it here.

http://www.funduc.com/search_replace.htm

Adam Stoller

Or on Unix: 'grep -n' and/or 'sed -e s/****/****/'
However, that will spit out the entire line in which the string was found and won't help in terms of providing the limited context for *each* occurrance found within a single string (which non-pretty-print DCRs are)

--fish
Senior Consultant, Quotient Inc.
http://www.quotient-inc.com

Edited by ghoti on 02/05/04 01:08 PM (server time).

nipper

Thanks fish, this helped a great deal, I was just going to grab an arbitrary number of characters before and after.

I cannot believe I get to critque your code but this:

my @text = ;

should have been
my @text = <IN>;

I am such an ingrate. :-)

Thanks again

Andy

Adam Stoller

Actually it was: my @text = <IN>; -- but I was doing it fast and forgot to protect it from the HTML formatting parsing of devnet (which means that when you use HTML formatting for your post you need to do: my @text = <<x>IN>;

Thanks for catching it though - in case anyone else was trying to use the code.

--fish
Senior Consultant, Quotient Inc.
http://www.quotient-inc.com