Shellscript Read XML attribute value

Question

We want to read XML attributes from an XML file. Example of file content is as below:

  
  <         


        

Answer 1

sed -n '/<property name="abc"/s/.*value="\(.*\)"[^\n]*/\1/p' file

Creates a hold pattern for the value then matches everything except for the newline to avoid printing the newline, it expects the value double quoted as per your example data.

E.g. 
<properties>
  <property name="abc" value="15"/>
  <property name="xyz" value="26"/>
</properties>

Output:
15

(Prior to edit: sed '/<property name="abc"/s/.*value="\(.*\)"[^\n]*/\1/' file)

Answer 2

You can use a proper XML parser like xmllint. If your version supports xpath, it will be very easy to grab specific values. If it doesn't support xpath, then you can use --shell option like so:

$ echo 'cat //properties/property[@name="abc"]/@value' | xmllint --shell myxml
/ >  -------
 value="15"
/ > 

You can then use awk or sed to format and extract desired field from output.

$ echo 'cat //properties/property[@name="abc"]/@value' | xmllint --shell myxmlfile | awk -F'[="]' '!/>/{print $(NF-1)}'
15

You can use command substitution to capture the output in a variable by saying:

$ myvar=$(echo 'cat //properties/property[@name="abc"]/@value' | xmllint --shell myxml | awk -F'[="]' '!/>/{print $(NF-1)}')
$ echo "$myvar"
15

Using anything else other than a xmlparser is prone to errors and will break easy.

Answer 3

quick and dirty

sed -n '/<Properties>/,\|</properties>| {
   s/ *<property name="xyz" value="\([^"]*\)"\/>/\1/p
   }'

no xml check and based on your sample so assume same structure (one property name per line, ...)

posix version (--posix for GNU sed)