Here is my json file. I would like to get string "c,d,e,f,g,h" using JQ query.

{
    "a1": [
      {
        "group_sourceId": "monday",
        "ids": ["a", "b"]
      },
      {
        "group_sourceId": "sunday",
        "ids": ["c", "d"]
      },
      {
        "group_sourceId": "sunday",
        "ids": ["e", "f"]
      },
      {
        "group_sourceId": "sunday",
        "ids": ["g", "h"]
      }
    ],
    "m1": [
      {
        "group_sourceId": "sunday"
      },
      {
        "group_sourceId": "sunday"
      }
    ]
}

This is what I tried.

cat /tmp/example.json | jq -r '.a1[] | select(.group_sourceId | startswith("sunday")) | .ids'

returns

[
  "c",
  "d"
]
[
  "e",
  "f"
]
[
  "g",
  "h"
]

then

cat /tmp/example.json | jq -r '.a1[] | select(.group_sourceId | startswith("sunday")) | .ids | join(",")'

c,d
e,f
g,h

How to get "c,d,e,f,g,h" as a output considering there could be any number of "ids" in "a1" object and any number of strings in "ids" array?

CodePudding user response：

Another approach using map. Also, you could use == to test for an exact match.

jq -r '.a1 | map(select(.group_sourceId == "sunday").ids[]) | join(",")'

c,d,e,f,g,h

Demo

CodePudding user response：

[.a1[] | select(.group_sourceId | startswith("sunday")) | .ids[]] | sort[]

produces the sorted listing. If you want the unique values, use unique instead of sort. If you want the output as a line of CSV, you could replace the final [] with | @csv.