Use Bash and jq to filter array of objects based on values in an inner array, plus x objects before and after matching objects?

Question

Similar to this problem except I am trying to include objects before and after the matched objects.

So for example I want to find all objects with type.name='pass', plus any object that is within X (say 2) of this object, either before or after.

This JSON:

    [
     {
       "class": "Something1",
       "type": {
         "name": "Foul"
       }
     },
     {
       "class": "Something2",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something3",
       "type": {
         "name": "Pass"
       }
     },
     {
       "class": "Something4",
       "type": {
         "name": "Pass"
       }
     },
     {
       "class": "Something5",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something6",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something7",
       "type": {
         "name": "Other"
       }
     },
     {
       "class": "Something8",
       "type": {
         "name": "Other"
       }
     },
     {
       "class": "Something9",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something10",
       "type": {
         "name": "Pass"
       }
     },
     {
       "class": "Something1",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something2",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something3",
       "type": {
         "name": "Carry"
       }
     },
     {
       "class": "Something4",
       "type": {
         "name": "Other"
       }
     },
     {
       "class": "Something5",
       "type": {
         "name": "Carry"
       }
     }
    ]

Would output a new JSON string:

    [
      {
        "class": "Something1",
        "type": {
          "name": "Foul"
        }
      },
      {
        "class": "Something2",
        "type": {
          "name": "Carry"
        }
      },
      {
        "class": "Something3",
        "type": {
          "name": "Pass"
        }
      },
      {
        "class": "Something4",
        "type": {
          "name": "Pass"
        }
      },
      {
        "class": "Something5",
        "type": {
          "name": "Carry"
        }
      },
      {
        "class": "Something6",
        "type": {
          "name": "Carry"
        }
      },
      {
        "class": "Something8",
        "type": {
          "name": "Other"
        }
      },
      {
        "class": "Something9",
        "type": {
          "name": "Carry"
        }
      },
      {
        "class": "Something10",
        "type": {
          "name": "Pass"
        }
      },
      {
        "class": "Something1",
        "type": {
          "name": "Carry"
        }
      },
      {
        "class": "Something2",
        "type": {
          "name": "Carry"
        }
      }
    ]

Or it could output the index of the above objects in a list which can then be used to search the original JSON.

I can filter by "type.name" thanks to the answer quoted above, but I could not work out how to include surrounding objects.

    $ passes=$(cat file.json | jq -c '[ .[] | select( .type.name | contains("Pass")) ]')

The files I am working with are 140,000+ lines long so efficiency is important.

Edit: Thanks @Gilles Quenot for fixing the code formatting.

Edit: Corrected errors in JSON and explained approach taken so far.

I edited the code for you, next time thanks to format your code like I have done, see edit — Gilles Quénot
– Gilles Quénot, Commented Jan 19, 2023 at 15:29
So did you write any code for this problem yet? As is, it sounds like you are asking us to do your work for you. But this is not how stackoverflow works; see How to Ask. — Robert
– Robert, Commented Jan 19, 2023 at 15:32
That's not valid JSON (missing commas and also what is "name::?) — knittl
– knittl, Commented Jan 19, 2023 at 15:36
Thank you for bearing with me and helping me write better questions. — Ed Rogers
– Ed Rogers, Commented Jan 19, 2023 at 16:29

knittl · Accepted Answer · 2023-01-19 17:42:34Z

Let's see …

Fix the input to be valid and well-formed JSON.
Get the index of each array element with to_entries.
Use map(select(…)) pattern to extract all keys (i.e. the index) which match your predicate.
"Pad" your extracted keys with adjacent keys: + range(3) - 1.
Store in variable.
Use map(select(…)) again on the entries to extract any item where the index matches one of the previously-extracted indices

Putting it all together:

to_entries
| map(select(.value.type.name=="Pass").key + range(3) - 1) as $keys
| map(select(.key|IN($keys[])).value)

range(3)-1 produces the stream -1,0,1 which means that all indices within range 1 will be checked. To check all indices with a max distance of 2, use range(5)-2.

Output:

[
  {
    "class": "Something2",
    "type": {
      "name": "Carry"
    }
  },
  {
    "class": "Something3",
    "type": {
      "name": "Pass"
    }
  },
  {
    "class": "Something4",
    "type": {
      "name": "Pass"
    }
  },
  {
    "class": "Something5",
    "type": {
      "name": "Carry"
    }
  },
  {
    "class": "Something9",
    "type": {
      "name": "Carry"
    }
  },
  {
    "class": "Something10",
    "type": {
      "name": "Pass"
    }
  },
  {
    "class": "Something1",
    "type": {
      "name": "Carry"
    }
  }
]

Collectives™ on Stack Overflow

Use Bash and jq to filter array of objects based on values in an inner array, plus x objects before and after matching objects?

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related