5 operations (1 bundle each) = 5 documents instead of 1 (Json Parser problem)

Hi everyone,

Couldn’t find any response to my problem anywhere, i’m stuck for about 4 days i’m going crazy.

I’m running 5 API sequences [Create Message + Run Assistant + Read Message] to an OpenAI Assistant using HTTP requests

Each run returns an output array where every object contains an unparsed JSON string that I want to extract and use in a document (Notion, Google Doc, etc.).

So first of all this is the output I get from an HTTP Request (it’s a JSON module because i’m working on a mock) :

If I zoom in, there is what I look for : datacontenttext.value

I only need data[1,3,5,7,9] because the other are not actual JSON strings. These are the input command on my OpenAI assistant so i don’t need them.

What I’ve tried

  • I use an Iterator on the output array to process each element
  • I filter so that only the actual JSON response strings are processed.
  • I then run each string through the JSON Parser module after the Iterator.

Main Issue :

  • The JSON Parser creates 5 separate operations with 1 bundle each which makes it impossible to aggregate all the data into a single mapping for Notion or Google Docs.

  • If I add an Array Aggregator after the JSON Parser, I get the same result: 5 operations, each with only 1 bundle.

  • If I put the Array Aggregator before the JSON Parser (immediately after the Iterator),
    and try to parse the array as a whole (e.g., on text.value for all items),
    the JSON Parser will only parse the first element of the array

What I want: Parse all my JSON strings and end up with a single bundle containing an array of all the parsed objects (by type: “situation”, “objectives”, etc.) for easy use in Notion/Google Doc modules.

Found the solution (eventually !!)

If this can help someone, here’s what I did :

  1. I added a text agregator to isolate my JSON string (with coma separator) :

  1. The JSON parser didn’t work with the text agregator output (whereas it was actually a JSON string)

  2. So I added this to my Json parser and Tadaaa :

Hope this helps
5.