[SOLVED] Parse XML tags with attribute with SED

adamzuber · 04-09-2012, 09:31 PM

Hi people, I've been searching all over the net for the correct regex/command, but cant find anything..

file.xml

Code:

<koko value="92029">
<arnab>5</arnab>
<kambing>3</kambing>
</xoxo>
<xoxo value="13245">
<kambing>2</kambing>
<kambing>3</kambing>
</xoxo>
<popo value="12345">
<kambing>2</kambing>
<kambing>3</kambing>
</popo>

I would like to extract only a spesific tag, including its tag attribute. Eg:

Code:

<xoxo value="13245">
<kambing>2</kambing>
<kambing>3</kambing>
</xoxo>

I tried egrep before with this regex, but nothing came up.

Code:

egrep '<xoxo[\s\S]*?/xoxo>' file.xml

Currently Im working for a one line solution like sed or egrep. I heard awk could do too but regex is too difficult for me to understand. Any hint on the solution is pretty much appreciated.

Thanks,
Adam.

jhwilliams · 04-09-2012, 10:01 PM

Awk would be a better next step, but I propose to use the xpath command. It's a wrapper around Perl's XML parser.

It assumes, though, that your document is valid XML. E.g.:

Code:

<?xml version="1.0"?>
<root>
  <koko value="92029">
    <arnab>5</arnab>
    <kambing>3</kambing>
  </koko>
  <xoxo value="13245">
    <kambing>2</kambing>
    <kambing>3</kambing>
  </xoxo>
  <popo value="12345">
    <kambing>2</kambing>
    <kambing>3</kambing>
  </popo>
</root>

Then:

Code:

jameson@yellow:~$ xpath -q -e '/root/xoxo' input.xml 
<xoxo value="13245">
  <kambing>2</kambing>
  <kambing>3</kambing>
</xoxo>

adamzuber · 04-09-2012, 10:22 PM

Hi jhwilliams, thanks for the reply. I manage to get what I want from the xpath, but, what if it is not a valid xml file? This is because I have few data that i have to merge.

Example:

Code:

cat file1.xml file2.xml > file3.xml

Assuming file1.xml and file2.xml is a valid xml file, now that i have file3.xml, which is not a valid xml file. I have tested with the 'xpath' but it breaks because of the invalid xml file and path. Any workaround?

adamzuber · 04-09-2012, 10:36 PM

Solved using sed.

Code:

sed '/<xoxo/,/<\/xoxo>/!d' notvalidxml.xml

Many thanks,
Adam.

firstfire · 04-09-2012, 10:37 PM

Removed