CODEWITHSUNDEEP
xmlbashshellawk
user3331975
user3331975

Reputation: 2815

print values from an xml output

I am working on a script, which returns output in xml format and wanted to print just the value of a particular attribute.

As an example, here is the output of the script :

~#] ./test.sh resource list --platform=centos

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<ResourcesResponse>
<Status>Success</Status>
<Resource id="11087" name="centos"

Now, I wanted to print only the Resource id which is 11087. When I used awk along with NR, it returns as below :

~#] ./test.sh resource list --platform=centos | awk 'NR==4{print $2}'

id="11087"

Could you please help know how to print only the value, i.e 11087

Upvotes: 1

Views: 390

Answers (4)

Idriss Neumann
Idriss Neumann

Reputation: 3838

Here is another solution using xmlstarlet with XPath query :

$ ./test.sh resource list --platform=centos|xmlstarlet sel -T -t -m '/ResourcesResponse/Resource/@id' -v '.' -n
11087
$ xmlstarlet sel -T -t -m '/ResourcesResponse/Resource/@id' -v '.' -n < <(./test.sh resource list --platform=centos)
11087

It's always better to use tools optimized for XML parsing (xmlstarter, xmllint, or more powerful shells languages ​​like perl, python, php cli mode, etc.).

Upvotes: 3

anishsane
anishsane

Reputation: 20980

grep variant:

grep -m1 -oP '(?<=id=")[0-9]*(?=")' file

Or with input piped from your command:

~#] ./test.sh resource list --platform=centos | grep -m1 -oP '(?<=id=")[0-9]*(?=")' 
11087

Explanation: Print only (-o) first match (-m1) of number ([0-9]*) prefixed by id=" ((?<=id=")) & followed by a " ((?=")).

Upvotes: 2

Jotne
Jotne

Reputation: 41460

This awk should give what you want.

awk -F\" 'NR==4{print $2}' file
11087

By setting Field Separator to " your data are in the second field.

To make sure you get correct id, I would have used:

awk -F\" '/Resource id/ {print $2}' file
11087

Upvotes: 2

John1024
John1024

Reputation: 113884

Using sed:

~#] ./test.sh resource list --platform=centos | sed -nr '4 s/.*id="([^"]+)".*/\1/p'
11087

Notes:

  • The -n option to sed tells it not print anything unless we explicitly ask it to.

  • The -r option to sed tells it to use extended regular expressions

  • The sed command 4 s/old/new/p tells it to operate on only on line 4 and, on that line, look for old and replace it with new and, only if that substitution happened, print the line.

  • In our case, the value of old is .*id="([^"]+)".*/. Since this starts with .* and ends with .*, it matches the whole line. It also captures the value of the id in match variable 1.

  • The value of new is simply \1 which is the value of the id.

Upvotes: 2

Related Questions