newbie perl question

dazdaz · 10-09-2012, 02:38 PM

Hi,

How can I write a perl regex to extract the data between : <string1>data</string1>

This may be spread over multiple lines too, depending upon the size of the data entry.

Quote:

Quote:

<string1>dataandyetmoredata
</string1>

Out of curiosity, how would you approach this in sed too.

ip_address · 10-10-2012, 01:00 AM

you can try this ..

# Input file

Code:

$ cat patterns1.txt 

<string1>data</string1>

# sed commands

Code:

$ sed -n '/<string1>/,/<\/string1>/p' patterns1.txt | sed -e 's/<string1>//g' -e 's/<\/string1>//g' | sed '/^$/d'

#Output

Code:

data

# Input file

Code:

cat patterns2.txt 

<string1>fdjfsfdjkjfdsal
fajsklsdjflk
</string1>

# sed commands

Code:

sed -n '/<string1>/,/<\/string1>/p' patterns2.txt | sed -e 's/<string1>//g' -e 's/<\/string1>//g' | sed '/^$/d'

#Output

Code:

fdjfsfdjkjfdsal
fajsklsdjflk

ntubski · 10-10-2012, 03:48 AM

Quote:

Originally Posted by dazdaz

How can I write a perl regex to extract the data between : <string1>data</string1>

That looks like XML, consider using an XML parser.

Quote:

Originally Posted by Jamie Zawinski

Some people, when confronted with a problem, think
“I know, I'll use regular expressions.” Now they have two problems.

sundialsvcs · 10-10-2012, 01:17 PM

Perl-related questions are probably better handled at http://www.perlmonks.org ...

markush · 10-10-2012, 01:45 PM

In order to parse tags over multiple lines you can undef the $/ variable.

Code:

undef $/ ;

Parsing can be done in the following way:

Code:

if ( $string =~ m/<string1>(.*?)<\/string1>/i ) {
    say $1 ;
}

where $1 holds the data between the tags.

Markus

bigearsbilly · 10-11-2012, 02:43 AM

If you are doing XML in perl I recommend http://xmltwig.org/

markush · 10-11-2012, 05:44 AM

In my opinion there are several cases when one doesn't need an XML-parser. If you only have one type of xml-files and only want to extract a few data where the tag-names are known, it is relatively easy in perl to do this with simple regex like I posted above.

I'd recommend to take a look at the sources of geolog http://geolog.sourceforge.net/geolog.html which helped me not to use a parser for a much too simple task.

Markus