Ansible get unique value in loop for register - ansible

Question

Hi friend below is my sample output from a regsiter name dataInfo
ok: [123.23.44.123] => {
"msg": [
{
"changed": true,
"item": [
{
"artifactName": "helloWorld.jar",
"status": "false"
},
"myGroup1"
],
"rc": 0,
"stderr_lines": [],
"stdout": "ok",
"stdout_lines": [
"ok"
]
},
{
"changed": true,
"item": [
{
"artifactName": "helloWorld.jar",
"status": "false"
},
"myGroup2"
],
"rc": 0,
"stderr_lines": [],
"stdout": "ok",
"stdout_lines": [
"ok"
]
}
Below is my code
- name: My ArtifactName Name
debug:
msg: "artifactName = {{ item.item[0].artifactName }}"
loop: "{{ dataInfo.results }}"
when: item.changed | bool == true and item.stdout == "ok"
How can i only display as below
artifactName = helloWorld.jar
rather than
artifactName = helloWorld.jar
artifactName = helloWorld.jar
I have shared my code as shown above. Please advice

Answer 1

As you are looping with dataInfo.results, the debug message will repeat for each "item". If you only want unique artifactName, then you can save it to a variable using json_query and choose to display only unique ones.
Something like this:
# Save list of artifactName found in dataInfo.results even with duplicates
- name: save artifact names
set_fact:
artifact_names: "{{ artifact_names | default([]) + [ item | json_query('item[].artifactName') ] }}"
loop: "{{ dataInfo.results }}"
# Show each item of artifact_names using unique filter to eliminate duplicates
- name: show unique artifact names
debug:
var: item
loop: "{{ artifact_names | flatten | unique }}"
Update:
Below set_fact:
artifact_names: "{{ artifact_names | default([]) + [ item.item[0].artifactName ] }}"
... will work if you get only 1 hash in item like this:
"item": [
{
"artifactName": "helloWorld.jar",
"status": "false"
},
],
But if it may have multiple hashes like below, it won't work:
"item": [
{
"artifactName": "helloWorld.jar",
"status": "false"
},
{
"artifactName": "someother.jar",
"status": "false"
},
...
],
If this is a possibility, then json_query should be used. It can be used exactly like the jq command.