find object id when key value in nested array has given string-CodePudding

Input json is:

[
  {
    "id": "00001",
    "profile": {
      "extensions": [
        {
          "name": "/somepath1...../AAAAAA"
        },
        {
          "name": "/somepath2...../xxxxxx"
        }
      ]
    }
  },
  {
    "id": "00002",
    "profile": {
      "extensions": [
        {
          "name": "/somepath1...../zzzzzz"
        },
        {
          "name": "/somepath2...../BBBBBB"
        }
      ]
    }
  },
  {
    "id": "00003",
    "profile": {
      "extensions": [
        {
          "name": "/somepath1...../yyyyyy"
        }
      ]
    }
  },
  {
    "id": "00004",
    "profile": {
      "extensions": []
    }
  }
]

I need to find id of object having .profile.extensions[].name having values containing string AAAAAA or BBBBBB.

Expected output for above json example would be:

00001 AAAAAA 
00002 BBBBBB

Because in the first one the is AAAAAA and in second BBBBBB

CodePudding user response：

You are probably looking for this filter pattern, which iterates over the array elements .[], filters for matching items using select and outputs the field id as raw text due to the -r flag.

jq -r '.[] | select( … ).id'

What the expression within your select filter should be, depends on how you define "having values containing [a] string".

If the search string has to match the last part of a slash-delimited list of substrings (as your sample data might suggest), you could split up the target string using / (or split), extract the last item and use IN to apply the test to a stream of search strings. (Demo)

IN(.profile.extensions[].name / "/" | last; "AAAAAA", "BBBBBB")

If the slashes should not be considered specially (i.e. a search string could also contain slashes) and if a search string simply appearing at the end of the target string is a match, then you can use endswith to test it and any to prevent multiple matches for the same id if several names within the same iterated object end with a given search string. (Demo)

[.profile.extensions[].name | endswith("AAAAAA", "BBBBBB")] | any

If, however, the search string is allowed to appear anywhere in the name, just change endswith to contains. (Demo)

[.profile.extensions[].name | contains("AAAAAA", "BBBBBB")] | any

Regarding your update which altered the output to include the matching search string, and your additional comment stating that the target string is indeed a slash-separated list, the approach now is as follows:

In order to test and output the search string, i.e. to use and to reference it again, we need to store it in a variable (using $s here), which has to happen outside the select filter in order to have its scope reach until the output reference.

jq -r '.[] | ("AAAAAA", "BBBBBB") as $s
  | select(IN(.profile.extensions[].name / "/" | last; $s))
  | "\(.id) \($s)"
'

00001 AAAAAA
00002 BBBBBB

Demo