XML parsing in KSH

Shell Programming and Scripting
1

Hi All,

Although I was able to find past XML parsing questions, as the questions were a little different than this, I had a pretty hard time adapting those answers to this scenario.

-> May I ask if anyone knows how to extract the IP addresses of the below "servers.xml" file into an array called IP[]?
(Hopefully using KSH in Solaris)
(note: there can be a varying # of server lines [sometimes zero] within this list, and will most likely be formatted on multiple lines, like below--Also note lines 1-2 and the last line below are guaranteed to be there.)

<?xml version="1.0" encoding="UTF-8"?>
<serverList xmlns="http://www.w3.org/1999/xhtml">
 <server hostname="server01" ip="192.168.1.101" loc="rm1" region="NY"/>
 <server hostname="server02" ip="192.168.1.117" loc="rm1" region="NY"/>
 <server hostname="server03" ip="192.168.1.154" loc="rm1" region="NY"/>
 <server hostname="server04" ip="192.168.1.159" loc="rm2" region="NY"/>
</serverList>

Thanks in advance...
CG

2

As you may have found, XML parsing can be very complex.

Each case should be treated uniquely.

As for the data sample you provided, try the code below:

sed -n '/ip="/s/.*ip="\(.*\)" loc.*/\1/p' input_xml
1 Like
3

To get the IPv4 addresses into an array:

$ set -A myarray `sed -n '/ip="/s/.*ip="\(.*\)" loc.*/\1/p' infile`
$ echo ${myarray[@]}
192.168.1.101 192.168.1.117 192.168.1.154 192.168.1.159
1 Like
4

xmllint is a very handy tool to work with XML files. Here is a foolproof way to get the ip addresss no matter how your XML file is formatted, even it is packed in one line.

echo "
setns a=http://www.w3.org/1999/xhtml
xpath /a:serverList/a:server[@ip]" | \
	xmllint --shell sample.xml | \
	awk '/ATTRIBUTE ip$/{getline;getline;split($0,a,"=");print a[2]}'

I blog about xmllint sometime ago because I cannot find any useful examples on the internet. See this Chi Hung Chan

1 Like
5

Thank you both Shell_Life & fpmurphy!

For those that need an XML parser to work similar to this scenario reading in the future, it turns out a perfect combination of the 2 suggestions above were needed. Thanks guys!

This is what works:

set -A myarray `sed -n '/ip="/s/.*ip="\(.*\)" ad.*/\1/p' /servers.xml`
echo ${myarray[@]}
192.168.1.101 192.168.1.117 192.168.1.154 192.168.1.159

Edit:
Chi Hung: Just before hitting submit, I noticed your post as well. Thank you! Hm, very interesting. I'll check this out now...especially since I plan to do a lot more XML parsing pretty soon! Thank you!

6

Sorry to contradict you but the script you came up with is not an XML parser in any shape, fashion or form. It is strictly a shell script. An XML parser is a completely different piece of software.