extract xml tag value from file where file name has spaces

Question

I have a file 'EXXARS - Physical Inventory Adjustment Report US PDF_es_ES.xlf'

file data:

<header>
 <prop-group name="ora_reconstruction">
    <prop prop-type="TemplateCode">XXPO_PHYS_INV_ADJ_ES_AS</prop>
    <prop prop-type="extractorVersion">10.1.3.4.2_1.5</prop>
 </prop-group>

I need extract the value between and result: I need this value in the variable XXPO_PHYS_INV_ADJ_ES_ES.

I have coomand which is not working

filename='ES - Physical Inventory Adjustment Report US PDF_es_ES.xlf'
sed -n 's:.*<prop prop-type="TemplateCode">\(.*\)</prop>.*:\1:p' filename > ${LOBCODE}
echo " --> ${LOBCODE}"

Answer 1

Assuming that the XML file is well formed:

<?xml version="1.0"?>
<header>
  <prop-group name="ora_reconstruction">
    <prop prop-type="TemplateCode">XXPO_PHYS_INV_ADJ_ES_AS</prop>
    <prop prop-type="extractorVersion">10.1.3.4.2_1.5</prop>
  </prop-group>
</header>

Using XMLStarlet:

xmlfile='ES - Physical Inventory Adjustment Report US PDF_es_ES.xlf'
LOBCODE=$( xml sel -t -v '//prop[@prop-type="TemplateCode"]' "$xmlfile" )

This uses XMLStarlet to extract the value of the prop node whose prop-type attribute is TemplateCode. The variable LOBCODE will get the value XXPO_PHYS_INV_ADJ_ES_AS with the above XML.

XMLStarlet may sometimes be installed as xmlstarlet rather than as xml.

---

There are two issues with your code:

1. You redirect the output of the sed command to a file whose name is given by $LOBCODE. If $LOBCODE is empty, then this redirection will fail. What I assume you wanted to do was to assign the output of sed to the variable LOBCODE. This is done with a command substitution as I've shown above. Note that sed is a poor choice of tool for parsing XML data.

2. You instruct sed to work with a file called filename. To have sed work on the value of the variable filename you will need to use $ in front of the variable name. Also, since the filename has spaces in it, you need to double quote the variable expansion (you should always do this, regardless). Hence you should use "$filename" (I've used a more descriptive variable name in my code above).

Answer 2

@Kusalananda's answer is excellent, but if you have xmllint instead, you can use this:

filename='ES - Physical Inventory Adjustment Report US PDF_es_ES.xlf'
LOBCODE=$(xmllint --xpath "header/prop-group[@name='ora_reconstruction']/prop[@prop-type='TemplateCode']/text()" "$filename")

Notes:

• I used the full path specification in case you only want the prop-group with the "ora-reconstruction" attribute. @Kusalananda's method assumes you want any prop element with an attribute of "TemplateCode": which method is preferable depends on your data.
• text() returns just the text inside the prop element (otherwise, xmllint returns the surrounding element as well.)

Example run of the important part:

$ xmllint --xpath "header/prop-group[@name='ora_reconstruction']/prop[@proptype='TemplateCode']/text()" "ES - Physical Inventory Adjustment Report US PDF_es_ES.xlf"
XXPO_PHYS_INV_ADJ_ES_AS

Answer 3

$ cat test.xml
<header>
 <prop-group name="ora_reconstruction">
    <prop prop-type="TemplateCode">XXPO_PHYS_INV_ADJ_ES_AS</prop>
    <prop prop-type="extractorVersion">10.1.3.4.2_1.5</prop>
 </prop-group>

$ awk -F"[<>]" '/TemplateCode/{print $3;exit}' test.xml
XXPO_PHYS_INV_ADJ_ES_AS