JQ Group_by or Unique not working for nested arrays

Trying to get unique entries from nested JSON arrays.

JSON Source

[
{"SIR":[{"PMD":[{"MMA":[{"DN":"One","SA":"1"},{"DN":"Two","SA":"2"}]}]}]},
{"SIR":[{"PMD":[{"MMA":[{"DN":"Two","SA":"2"},{"DN":"Four","SA":"4"}]}]}]},
{"SIR":[{"PMD":[{"MMA":[{"DN":"Three","SA":"3"},{"DN":"Five","SA":"5"}]}]}]}
]

Filter producing non-unique results:

'.[].SIR.PMD.MMA 
    | [.DN, .SA]
    | @csv'

Current Output (CSV

One,1
Two,2
Two,2
Three,3
Four,4
Five,5

Desired Output (CSV)

One,1
Two,2
Three,3
Four,4
Five,5

Collect the items to be filtered into an array, and use either unique alone to compare the whole objects, or unique_by on any filter combination, e.g. .DN):

jq -r 'map(.SIR[].PMD[].MMA[]) | unique[] | [.DN, .SA] | @csv'

jq -r 'map(.SIR[].PMD[].MMA[]) | unique_by(.DN)[] | [.DN, .SA] | @csv'

Note that the output will be sorted, in the first case by the whole object, in the latter by the filter provided. To achieve the numeric ordering of your sample output, use unique_by on either just (.SA|tonumber), or first (.SA|tonumber) then .DN. The conversion to numbers only affects the sorting, not the output:

jq -r 'map(.SIR[].PMD[].MMA[]) | unique_by(.SA|tonumber)[] | [.DN, .SA] | @csv'

jq -r 'map(.SIR[].PMD[].MMA[]) | unique_by([(.SA|tonumber), .DN])[] | [.DN, .SA] | @csv'

"One","1"
"Two","2"
"Three","3"
"Four","4"
"Five","5"