Usually XML processing is already implemented in programming languages. Several of them have numerous API for that. However some of these API are awkward and cumbersome. Recently it was necessary to parse xml file from bash script so I create my own parser. Basically what I create it is not inversion but small and it works for me, so I want to share it with you. The xml file sample is:
|
<?xml version="1.0" encoding="UTF-8"?> <Customer 1> <First name>Jesus</First name> <Second name>Christ</Second name> </Customer 1> <Customer 2> <First name>Mary</First name> <Second name>Magdalene</Second name> </Customer 2> </xml> |
The get_xml_node_value.sh script itself is:
|
#! /bin/bash get_xml_node_value() { startline=$(grep -n "<$2" $1 | cut -d':' -f1) lastline=$(grep -n "<\/$2>" $1 | cut -d':' -f1) value=$(sed -n $startline,$lastline'p' $1) value=$(echo $value | sed -e "s/^.*<$3/<$3/" | sed -e "s/<$3>//g" | sed -e "s/<\/$3>//g" | awk -F"<" '{print $1}' | sed -e 's/[[:space:]]*$//' ) } if [ $# -eq 3 ]; then if [ -f $1 ]; then get_xml_node_value $1 "$2" "$3" echo $value else echo "File $1 not found" fi else echo "3 arguments are required: xml file name, section name, node name" echo "Example:" echo "./get_xml_node_value abc.xml \"Customer 1\" \"First name\"" fi |
Script testing results:
|
./get_xml_node_value.sh abc.xml "Customer 1" "First name" Jesus # ./get_xml_node_value.sh abc.xml "Customer 2" "Second name" Magdalene |
The node names in arguments are case sencitive.
Thanks a lot for sharing!
Parsing the XML with a pure bash is not the easiest way. There are fine tools like xmllint, but sometimes we meet limitations about the installations on the servers we have to work with.
You have done a good approach that works even for more complicated XML structure as it was presented.