bash extract xml attribute value using xmllint

Question

I am extracting an attribute value from xml file, but I get an error. I'd like to extract the value for key="qua" in the firstpart element. Here is my script, but below you find the errors:

#!/bin/bash

myfile=$1

myvar=$(echo 'cat //firstpart/step/category/id/info[@key="qua"]/@value' | xmllint --xpath "$myfile" | awk -F'[="]' '!/>/{print $(NF-1)}')

echo "$myvar"

how my xml file looks like:

<?xml version='1.0' encoding='UTF-8'?>
<firstpart>
    <step name="Home">    
        <category name="one">
            <id name="tools">
                <info key="qua" value="1"/>        
            </id>
        </category>
    </step>
    <step name="Contact">    
        <category name="two">
            <id name="tools">
                <info key="qua" value="2"/>        
            </id>
        </category>
    </step>
    ...
</firstpart>
<secondpart>
    <step name="office">    
        <category name="one">
            <id name="tools">
                <info key="qua" value="100"/>        
            </id>
        </category>
    </step>
    <step name="Contact">    
        <category name="two">
            <id name="tools">
                <info key="qua" value="200"/>        
            </id>
        </category>
    </step>
    ...
</secondpart>

the errors I get:

awk: run time error: negative field index $-1
    FILENAME="-" FNR=71 NR=71

./mybash.sh: line 3: $: command not found
./mybash.sh: line 4: $: command not found

Solution 1

It looks like you call xmllint in wrong way.

xmllint --xpath '//firstpart/step/category/id/info[@key="qua"]/@value' FILE.xml

Result:

value="1" value="2"

Complete script:

#!/bin/bash

str=$(xmllint --xpath '//firstpart/step/category/id/info[@key="qua"]/@value' $1)

entries=($(echo ${str}))
for entry in "${entries[@]}"; do
    result=$(echo $entry | awk -F'[="]' '!/>/{print $(NF-1)}')
    echo "result: $result"
done

May be, it's not better solution, but at least it works :)

Solution 2

Get value of attribute with xmllint:

xmllint --xpath 'string(//firstpart/step[1]/category/id/info/@value)' file.xml

Output:

1

Comments

[Ed Morton]

Have you tried saving the xmllint output to a file and running awk on that? I assume not since your problem will be extremely obvious if you do.

[passion]

@EdMorton how can i do it? could you provide an answer please?

[choroba]

Instead of piping to awk, redirect to a file. Then run awk on that file and get the same error. Then inspect the file to see the problem.

[Cyrus]

Your file is not valid XML.

[passion]

@Cyrus why is it not a valid one? i have excluded the tail of the file and included three dots in the middle to say that they are other stuff there...

[Cyrus]

What is your desired output? key="qua" or qua?

[Cyrus]

@passion: An XML has one root node: <foo><firstpart></firstpart><secondpart></secondpart></foo>

[passion]

how can i get the value only and not the ` value="1"` ? on the the number

[passion]

I like this solution but since Bor laze was first to answer, so I accept that one as the answer to the question.