Im freaking out bruned operations like hell. I want to update specific rows in Airtable after they got processed from an GPT. I managed to get the data to the Airtable module and now Im stuck for days. all I try is not working please find attached the blueprint. Thanks for your help.
blueprint (2).json (45.8 KB)
Hello @Andreas_P
With the blueprint we can’t see everything, but from what I saw, apparently you are trying to map a collection “Deliverable” in the “Deliverable” field of Airtable. So Make doesn’t know how to transform that to a text.
Can you double-check what this field contains, and see what you want to get from it before you send it to Airtable? You should be able to map the specific field inside the Deliverable, or use a function to extract/build the relevant data.
If you are still stuck, can you please send us a screenshot of the output of Parse JSON (where we could see what contains Deliverable), and tell us what the field Deliverable in Airtable should look like?
Benjamin
HI Benjamin,
Thanks for your reply.
Here is the input :
[
{
“json”: " [\n {\n "ID": "rec80bsSNdvm2wvDP",\n "Deliverable": {\n "Prediction": 0.5,\n "Reason": "Dollar weakness brings relief."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": 0.1,\n "Reason": "Positive news for Harris."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": -0.3,\n "Reason": "Eurozone factory activity weak."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": -0.7,\n "Reason": "Hedge funds betting against banks."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": -0.2,\n "Reason": "Brazil court to vote on X ban."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": -0.4,\n "Reason": "Eurozone factory activity stagnant."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": -0.8,\n "Reason": "German manufacturing downturn accelerates."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": 0.2,\n "Reason": "Alstom sells signaling business."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": -0.1,\n "Reason": "PRS REIT questions shareholders."\n }\n },\n {\n "ID": "rec91q8296n204z4Y",\n "Deliverable": {\n "Prediction": 0.6,\n "Reason": "Blackstone nears AirTrunk acquisition."\n }\n }\n]\n "
}
]
and the Output :
[
{
“ID”: “rec80bsSNdvm2wvDP”,
“Deliverable”: {
“Prediction”: 0.5,
“Reason”: “Dollar weakness brings relief.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: 0.1,
“Reason”: “Positive news for Harris.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: -0.3,
“Reason”: “Eurozone factory activity weak.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: -0.7,
“Reason”: “Hedge funds betting against banks.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: -0.2,
“Reason”: “Brazil court to vote on X ban.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: -0.4,
“Reason”: “Eurozone factory activity stagnant.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: -0.8,
“Reason”: “German manufacturing downturn accelerates.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: 0.2,
“Reason”: “Alstom sells signaling business.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: -0.1,
“Reason”: “PRS REIT questions shareholders.”
}
},
{
“ID”: “rec91q8296n204z4Y”,
“Deliverable”: {
“Prediction”: 0.6,
“Reason”: “Blackstone nears AirTrunk acquisition.”
}
}
]
Screenshot :
I really tried to update the Airtable but the best I got was a update always in the same row
Airtable screenshot :
Thank you a lot !
1 Like
Hey @Andreas_P
Thanks for the update!
Did it work at the end or are you still stuck? Sorry, I didn’t understand 
still stuck.
I managed to get an update into one row but that`s it.
Hello.
From what I see, the IDs are all the same, except for the first bundle.
Where do you generate these ids, aren’t they supposed to be different?
Benjamin
these ID`s are the one from the Airtable.
What I want to do is get content from an airtable process in an GPT and insert the result into the same row at another colum.
So I thought to get all data out of the airtable. convert it to one string, to avoid rate limitation at the gpt . Tell the GPT to seperate the answers and output a JSON to work with and then based on the ID of the Airtable row in the JSON structure insert the prediction and content into the airtable.
@Andreas_P thanks! It’s clearer and clearer for me 
So you want to update a single row but to have all predictions the GPT returned?
The way you built the scenario, Make will loop the list of predictions and update the raw for each. The values are replaced, so you should see only the last value.
Do you have a column in Airtable that accepts an Array so that you can push all predictions for the row?
If not, and if you want a list of predictions separated by a comma, you can definitely build it with a Text Aggregator.
I attach you a little sample scenario I built to show you the possiblities:
Aggregate Predictions and Reasons.json (13.9 KB)
I hope it helps
Benjamin