How to Parse the XML response of a URI in ansible - ansible

Question

There is a code which call the web services of an application.
- uri:
url: http://10.164.52.61:8080/ems/v74/ws/customer.ws?customerRefId=f4XXXb15d69c3
method: GET
content_as_json: true
password: admin
user: admin
validate_certs: no
return_content: yes
HEADER_Cookie: "{{login.set_cookie}}"
register: customerId
- debug:
var: customerId.content
The sample response is
"customerId.content": "<listResponse type=\"customer\" count=\"1\"><instance customerId=\"28\" name=\"abc\" customerRefId=\"xyz\" refId1=\"12\" type=\"org\" enabled=\"true\" phone=\"\" fax=\"\" billingZip=\"\" billingAddress=\"\" billingCity=\"\" billingCountry=\"\" billingState=\"\" vendor=\"1\" defaultEmail=\"test\" defaultContactName=\"test\"/></listResponse>"
I want to access the list response in my next block of code. Like i need just the value of "customerId". How this can be achieved using anisble

Answer 1

This is similar to sebaszw's answer, however doesn't require the xml response to be written to a file, instead storing it in a variable.
- uri:
url: http://10.164.52.61:8080/ems/v74/ws/customer.ws?customerRefId=f4XXXb15d69c3
method: GET
content_as_json: true
password: admin
user: admin
validate_certs: no
return_content: yes
HEADER_Cookie: "{{login.set_cookie}}"
register: customerId
- xml:
xmlstring: "{{customerId.content}}"
xpath: /listResponse/instance
content: attribute
register: instance_attributes
- debug:
var: instance_attributes.matches.0.instance.customerId

Answer 2

There is no XML support in Ansible out of the box.
You can use regex filters to make some simple searches in your XML data.
In your example:
- debug: msg="{{ customerId.content | regex_findall('customerId=\"(\d+)\"') }}"
Will search for customerId=\"<number>\" strings and return a list of numbers in quotes.
Update: fully working example:
- hosts: localhost
gather_facts: no
vars:
myvar: "<listResponse type=\"customer\" count=\"1\"><instance customerId=\"28\" name=\"abc\" customerRefId=\"xyz\" refId1=\"12\" type=\"org\" enabled=\"true\" phone=\"\" fax=\"\" billingZip=\"\" billingAddress=\"\" billingCity=\"\" billingCountry=\"\" billingState=\"\" vendor=\"1\" defaultEmail=\"test\" defaultContactName=\"test\"/></listResponse>"
tasks:
- debug: msg="{{ myvar | regex_findall('customerId=\"(\d+)\"') }}"
I use ansible 2.1.1.0.

Answer 3

From Ansible 2.4 you can do this:
- get_url:
url: http://10.164.52.61:8080/ems/v74/ws/customer.ws?customerRefId=f4XXXb15d69c3
dest: ./response.xml
url_password: admin
url_user: admin
validate_certs: no
headers:
Cookie: "{{login.set_cookie}}"
- xml:
path: ./response.xml
xpath: /listResponse/instance
content: attribute
register: instance_attributes
- debug:
var: instance_attributes.matches.0.instance.customerId

Answer 4

I just added XML parsing support in the uri module because I needed it too.
https://github.com/ansible/ansible/pull/53045
Just like the JSON support, it will return an 'xml' key with a dictionary consisting of the XML content for the convenience of accessing the data in the payload.
Your example would look like this:
- uri:
url: http://10.164.52.61:8080/ems/v74/ws/customer.ws?customerRefId=f4XXXb15d69c3
method: GET
password: admin
user: admin
validate_certs: no
HEADER_Cookie: "{{login.set_cookie}}"
register: customerId
- debug:
var: customerId.xml.listResponse.instance['#customerId']
The output in customerId.xml would be:
{
'listResponse': {
'#type': 'customer',
'#count', '1',
'instance': {
'#customerId': '28',
'#name': 'abc'
'#customerRefId': 'xyz',
'#refId1': '12',
'#type': 'org',
'#enabled': 'true',
'#phone': '',
'#fax': '',
'#billingZip': '',
'#billingAddress': '',
'#billingCity': '',
'#billingCountry': '',
'#billingState': '',
'#vendor': '1',
'#defaultEmail': 'test',
'#defaultContactName': 'test'
}
}
}

Answer 5

Its worth looking at ansible-xml for this. It's not part of Ansible but its a super useful module for parsing xml, it uses the lxml Python library under the hood.

Answer 6

As this was reasonably complex getting this to work I thought I'd post my Ansible - XML parameters for parsing a SOAP response:
<S:Envelope xmlns:S="http://schemas.xmlsoap.org/soap/envelope/" xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/">
<SOAP-ENV:Header/>
<S:Body xmlns:ns2="http://ws.coverity.com/v9">
<ns2:getLdapServerDomainsResponse>
<return>
<name>soapui_test</name>
</return>
</ns2:getLdapServerDomainsResponse>
</S:Body>
</S:Envelope>
The trick was getting namespaces and xpath right. Also as I wanted the actual name I needed content: text
- name: Extract the name from the SOAP
xml:
xmlstring: "{{ ldap_response.content }}"
xpath: "/S:Envelope/S:Body/ns2:getLdapServerDomainsResponse/return/name"
namespaces:
S: "http://schemas.xmlsoap.org/soap/envelope/"
SOAP-ENV: "http://schemas.xmlsoap.org/soap/envelope/"
ns2: "http://ws.coverity.com/v9"
content: text
register: ldap_name
I still don't understand why the xpath section at the end doesn't work with .../ns2:return/ns2:name as the Ansible - XML docs example suggests. The above fills ldap_name with (simplified):
"ldap_name": {
"matches": [
{
"name": "soapui_test"
}
],
}