perl replace script for xml

ruffles · 06-16-2008, 12:34 PM

I'm trying to replace ocurrence of xml fragment in a large xml file.

replacement from:

Code:

$what= <<ENDHTML;

<image type="filmposter" id="$rd">
<imagefile type="tiny">
<filetype>jpg</filetype>
<filedate>DATA_ZDJ</filedate>
<filesize>ROZMIAR</filesize>
<location>images/1320521STARTtiny1.jpg</location>
<width>SZEROKOSC</width>
<height>WYSOKOSC</height>
</imagefile>
<imagefile type="thumb">
<filetype>jpg</filetype>
<filedate>DATA_ZDJ</filedate>
<filesize>ROZMIAR</filesize>
<location>images/1320521STARTtiny1.jpg</location>
<width>SZEROKOSC</width>
<height>WYSOKOSC</height>
</imagefile>
<imagefile type="photo">
<filetype>jpg</filetype>
<filedate>DATA_ZDJ</filedate>
<filesize>ROZMIAR</filesize>
<location>images/1320521STARTphoto.jpg</location>
<width>SZEROKOSC</width>
<height>WYSOKOSC</height>
</imagefile>
</image>
ENDHTML

to:

Code:

$whatfor= <<ENDHTML;

<image type="filmposter" id="$rd">
<imagefile type="tiny">
<filetype>jpg</filetype>
<filedate>$ARGV[1]</filedate>
<filesize>$ARGV[2]</filesize>
<location>images/1320521STARTtiny1.jpg</location>
<width>$ARGV[3]</width>
<height>$ARGV[4]</height>
</imagefile>
<imagefile type="thumb">
<filetype>jpg</filetype>
<filedate>$ARGV[1]</filedate>
<filesize>$ARGV[2]</filesize>
<location>images/1320521STARTtiny1.jpg</location>
<width>$ARGV[3]</width>
<height>$ARGV[4]</height>
</imagefile>
<imagefile type="photo">
<filetype>jpg</filetype>
<filedate>$ARGV[1]</filedate>
<filesize>$ARGV[5]</filesize>
<location>images/1320521STARTphoto.jpg</location>
<width>$ARGV[6]</width>
<height>$ARGV[7]</height>
</imagefile>
</image>
ENDHTML

So far I've tried:

Code:

#! /usr/bin/perl -w

##insert above variables here##

$file = '/path/to/file/file.xml';
$file =~ s/$what/$whatfor/g

and

Code:

#! /usr/bin/perl -w

##insert above variables here##

perl -pi -e "s/$what/$whatfor/e" '/path/to/file/file.xml'

None of them worked, of course, since this is my first perl script.

Could you point me in the right direction?

osor · 06-16-2008, 04:33 PM

If you want to pursue the first usage, you will have to open and overwrite the file yourself. If you want to use the second usage, just put the entire command as an argument to system.

ruffles · 06-16-2008, 04:42 PM

As for the first one, well, I actually don't really know how to do that. Did some googling, found a few sample scripts, but they didn't work for me.
As for the second, I tried it and it worked ( with simple replacements ), but i don't know how to pass a variable from shell (sh) to perl command ( it didn't work when I just placed the whole string in the command line )

sundialsvcs · 06-16-2008, 10:26 PM

Perl has many CPAN-modules for manipulating XML. These include libraries specifically designed to do XML search-and-replace, e.g. through XPath-expressions.

Check out, for example, XML::Twig.

A CPAN-search for XML will produce a huge number of entries. This is a great example of a situation where "automatically dropping into procedural-programming mode... 'how do I write a program to do this?'" will instantly be the wrong approach. The answer is "you don't." The program has already been written. Searching through and transforming an XML structure can be done declaratively, using the facilities provided by the CPAN modules.

ruffles · 06-17-2008, 12:17 AM

Problem is that i don't have control over modules installed on the server and there are no xml related modules already installed.
Also, time is an issue here as well.
I just need to replace one with the other, and sed is throwing errors about suffix larger than 512

ghostdog74 · 06-17-2008, 12:39 AM

here's an awk script. you can convert that to Perl.

Code:

awk '
/filedate|filesize|width|height/ {
  gsub(/>.*</,">$ARGV["++counter "]<")
  print;next
 }1' file

ruffles · 06-17-2008, 01:58 AM

this script should find occurence of string depending on $rd variable
<image type="filmposter" id="$rd">
it's called from loop within .sh script

I also tried this:

open(FILE,"</path/to/file/file.xml");
while(<FILE>){
$_=~s/$what/$whatfor/;
push @new_file,$_;
}
close(FILE);
open(FILE,">/path/to/file/file1.xml");
for $i(0..$#new_file){
print FILE $new_file[$i];
}
close(FILE);

but that didn't work either

ruffles · 06-17-2008, 05:39 AM

I solved this problem and I'll try to post the answer for later generations later on

ruffles · 06-17-2008, 05:30 PM

Well, here you go, maybe it'll help someone:

file change.pl:

Code:

#!/usr/local/bin/perl

$what="<image type=\"filmposter\" id=\"" . $ARGV[0] . "_pic\"\/>";

while (<STDIN>) {
     s/$what/<image type="filmposter" id="$ARGV[0]_pic">\n<imagefile type="tiny">\n<filedate>$ARGV[1]<\/filedate>~AND_SO_ON~<\/image>/;
        print;

and
chmod a+x change.pl

within sh script:

Code:

cd /dir/with/change.pl/script/
cat /path/to/file/file.xml | ./change.pl $a $d > /path/to/file/file1.xml
mv /path/to/file/file1.xml /path/to/file/file.xml

(if you want to change all occurences of a string, change
<\/image>/;
to
<\/image>/g;
in change.pl)

Thanks for your help.