Adding a new Field in array elements

Hi,

We were struggling to copy and add a new element in case of array,

Example : Input :

{
    "customer": [
        {
            "locations": [
                {
                                        "name" : "Location1",
                    "id": 1234
                },
                {
                                        "name" : "Location2",
                    "id": 4567
                }
            ]
        }
    ]
}

Expected Output :

{
    "customer": [
        {
            "locations": [
                {
                                        "name" : "Location1",
                    "id": 1234,
                                       "customId" : 1234,
                },
                {
                                        "name" : "Location2",
                    "id": 4567,
                                        "customId" : 4567   
                }
            ]
        }
    ]
}

Here “customId” is a field which is a copy of “id” field in the array elements. We Tried the following spec:

[
  {
    "operation": "shift",
    "spec": {
      "customer": {
        "*": {
          "locations": {
            "*": {
              "id": "customer[&1].id",
              "name": "customer[&1].name",
              "id": "customer[&1].customId"
            }
          }
        }
      }
    }
  }
]

But couldn’t achieve it. We were expecting not to touch other fields in spec(as the input json will be a very huge file) and achieve the final output only adding the info about the “customId” field in the spec.

Any help is very much appreciated.

[mhubbard]

There’s a few issues with your spec, however the main one is that you have a json object that has duplicate keys of “id”, when jackson tries to transform your spec into a java object to be fed into the jolt engine it writes the first id key with a value of “customer[&1].id” then overwrites it with “customer[&1].customId”.

Luckily what you’re trying to do can be done with a the default operation which will avoid the duplicate keys

[
  {
    "operation": "modify-default-beta", 
    "spec": {
      "customer": {
        "*": {
          "locations": {
            "*": {
              "customId": "@(1,id)"
            }
          }
        }
      }
    }
  }
]

Just for future reference your shift operation wasn’t fully recreating your tree, you needed to add the locations path to the right hand side as well, and traverse 2 levels higher for the right customer array index. If the duplicate key wasn’t an issue would would be able to do the shift like this:

[
  {
    "operation": "shift", 
    "spec": {
      "customer": {
        "*": {
          "locations": {
            "*": {
              "id": "customer[&3].locations[&1].id",
              "id": "customer[&3].locations[&1].customId", 
              "name": "customer[&3].locations[&1].name"
            }
          }
        }
      }
    }
  }
]

[mhubbard]

I too have spent some time on errors like that, that request should probably go in it’s own issue though. It’s not too hard to fix, just remove the whitespace before passing to the traversal builder in each *LeafSpec.java

For instance line 62 in ShiftrLeafSpec.java becomes

writers = Arrays.asList( TRAVERSAL_BUILDER.build( ((String)rhs).replaceAll("\\s+","") ) );

And lines 137 and 140 become the following

return FunctionArg.forContext( TRAVERSAL_BUILDER.build( arg.substring( 1 ).replaceAll("\\s+","") ) );

return FunctionArg.forSelf( TRAVERSAL_BUILDER.build( arg.replaceAll("\\s+","") ) );

This is just for the right hand side though, the keys/lhs requires similar changes to parent class of the *LeafSpec classes.