I’m trying to extract info on certain specific dependencies from the output of DTrack’s /v1/project endpoint.

The API returns data in the format:

[{
    "name": "project name",
    "directDependencies": <escaped JSON string in which I need to search for "com.example.commons/[email protected]" and similar strings>
}, ...]

Here’s what I got so far:

jq '.[] | select(.classifier=="APPLICATION") | {
      id: .name,
      dep1: ((.directDependencies // "") | capture(["com\.example\.commons/dep1@(?<version>[0-9.]+)"])),
      dep2: ((.directDependencies // "") | capture(["com\.example\.commons/dep2@(?<version>[0-9.]+)"])),
      dep3: ((.directDependencies // "") | capture(["com\.example\.commons/dep3@(?<version>[0-9.]+)"])),
      dep4: ((.directDependencies // "") | capture(["com\.example\.commons/dep4@(?<version>[0-9.]+)"]))
    }'

Generated output looks fine:

{
  "id": "project1",
  "dep1": {
    "version": "0.102.0"
  },
  "dep2": {
    "version": "0.102.0"
  },
  "dep3": {
    "version": "0.102.0"
  },
  "dep4": {
    "version": "0.102.0"
  }
}
... // more elements

The problem is, the output only contains those of the input elements for which all four captures matched something.

I would like the output to contain all the elements of the input list, even if some of the captures fail to match. In other words, I would expect elements in the output like:

{
    "id": "project1",
    "dep1": { "version": "0.1.0" },
    // dep2 didn't match
    // dep3 didn't match
    "dep4": { "version": "0.3.0" }
}

I couldn’t find such an option in the docs. Any ideas how to solve this?