AEP Datastream Data Mapping/Data Prep: Help needed to transform objects in an array using calculated fields for analytics list variables

Forum|Forum|1 year ago
September 13, 2024
3 replies
1586 views

I'm trying to use Calculated Fields in Data Mapping to transform objects in an input array that look like [{"key1":"string1","key2:"string2"},{"key1":"string1","key2:"string2"}] to an output array of objects that look like [{"value":"string1=string2"},{"value":"string1=string2"}], but I can't figure out the format needed for the input Source Fields, the combination of calculated fields needed, or the format needed for the output Target Field.

The objects in the array are always a pair of key:string, but the array is a dynamic size, sometimes containing dozens of {"key1":"string1","key2:"string2"} objects, so it's not feasible to individually reference the fields in each object in the input array entry via [0], [1], [2], etc. I tried using data.page.facetRefinement[*].name but the Calculated Field tool won't accept it.

Example input JSON:

{
  "data": {
    "page": {
      "facetRefinement": [
        {
          "name": "Document Type",
          "value": "PDF"
        },
        {
          "name": "Category",
          "value": "Important Stuff"
        }
      ]
    },
    "eventInfo": {
      "component": [
        {
          "type": "Image",
          "name": "Hero"
        }, {
          "type": "List",
          "name": "What's new!"
        }, {
          "type": "Carousel",
          "name": "Featured Things"
        }
      ]
    }
  }
}

Desired output (note: desired output format corrected based on @kautuk_sahni's reply):

{
  "_experience": {
    "analytics": {
      "customDimensions": {
        "lists": {
          "list1": [
            {"value": "Document Type=PDF"},
            {"value": "Category=Important Stuff"}
          ],
          "list2": [
            {"value": "Image=Hero"},
            {"value": "List=What's new!"},
            {"value": "Carousel=Featured Things"}
          ]
        }
      }
    }
  }
}

Anyone have any ideas?

Thanks!
-Phil

Experience Platform

This post is no longer active and is closed to new replies. Need help? Start a new post to ask your question.

bjoern__koth

Community Advisor and Adobe Champion

Hi @philschmidt

I saw you asked exactly the same question here, without answer yet. Can you give a quick update on what you have tried so far?

let's take one step back and check your output format. Pretty sure your array shall not contain one object with two "value" attributes.

          "list1": [
            {
              "value": "Document Type=PDF",
              "value": "Category=Important Stuff"
            }
          ],

https://experienceleague.adobe.com/en/docs/analytics/implementation/vars/page-vars/list

lists are arrays of individual objects in your array instead

          "list1": {
            "list": [
              {
                "value": "Example value 1"
              },
              {
                "value": "Example value 2"
              },
              {
                "value": "Example value 3"
              }
            ]
          }

The inline documentation in the UX is really not helpful and is missing out explaining half of the function parameters. Not really surprised, Adobe...

After looking at the documentation, things seem to get more clear.

So, my best best would be to generate the whole array

to_array(true, to_object("value", concat(data.page.facetRefinement[*].name, "=", data.page.facetRefinement[*].value)))

to_array(true, to_object("value", concat(data.eventInfo.component[*].type, "=", data.eventInfo.component[*].name)))

and map it

_experience.analytics.customDimensions.lists.list1.list[*]

BUT: the UX validation/preview fails if asterisks are used

Alternatively, do try without the asterisks and see what the AEP Debugger logs show for mapping events.
The preview shows only a single item from the array though. So, no idea whether this is an error in the UX or if it by design only takes one item if no index is defined.

add_to_array([], to_object("value", concat(data.page.facetRefinement.name, "=", data.page.facetRefinement.value)))

add_to_array([], to_object("value", concat(data.eventInfo.component.type, "=", data.eventInfo.component.name)))

and map them to your list.

_experience.analytics.customDimensions.lists.list1.list

Cheers from Switzerland!

philschmidtAuthor

Level 2

Thank you for your reply!

@bjoern__koth wrote:
I saw you asked exactly the same question here, without answer yet. Can you give a quick update on what you have tried so far?

That was someone else's question that I commented on. @jinx0011's question was similar, but it was not exactly the same question. The main reason I asked this question separate from @jinx0011's question is because I did not want to presume the solution required the "Ability to use array notation ([*]) in calculated field". I also wanted to provide a set of input and output JSON examples so if people wanted they could paste the JSON into Data Mapping to define the incoming data and test/validate potential solutions.

Here's everything I that I tried (among calculated fields that Data Mapping allowed me to save without throwing an error), and their results:

============= First attempt =============

add_to_array(to_array(true,""), concat(data.eventInfo.component.type, "=", data.eventInfo.component.name)) -> _experience.analytics.customDimensions.lists.list2.list[*].value

=== Result (from analytics.hit in AEP Assurance session) ===

mvvar2: [Ljava.lang.Object;@7f24998d|[Ljava.lang.Object;@27082dd5=[Ljava.lang.Object;@5458d032