Hi everyone, I’m trying to aggregate multiple ChatGPT responses into a single string using the Text Aggregator, but it’s not working as expected.
Scenario:
I extract a long text and split it into parts using a Text Parser.
I use an Iterator to send each part to ChatGPT, which returns a paraphrased version for each block.
ChatGPT generates 6 separate bundles, one for each part of the text.
I want to merge these 6 bundles into a single text using Text Aggregator and then update Google Sheets.
Problem:
Even though Text Aggregator is set up correctly (Source Module: ChatGPT, Row Separator: \n\n), it seems that the bundles are not actually merging.
When updating Google Sheets, only the last bundle is written, overwriting the others.
How can I aggregate the text and update Google Sheets only after? Where am I going wrong?
I have also tried using the Array Aggregator, but I am still facing the same issue. The text is not being aggregated correctly, and when updating Google Sheets, only the last bundle is written, overwriting the previous ones.
How can I properly aggregate the text and update Google Sheets only after the full text has been combined? Where am I going wrong?"
Hi, blueprint.json (68.8 KB)
here is the blueprint, I have tried every way with both Array Aggregator and Text Aggregator but after all these tests I think these modules are meant for extracting records and not for extracting and aggregating data from operations like mine.
Here is what I think it’s happening
The aggregator is outputting only 1 thing (an array). So the next module (update a row) is only performing 1 operation= updating only 1 row. Probably you are putting in column O only the first element of the array.
Might be 2 solutions
After the array aggregator put an array iterator (or a repeater) so that you can insert the N elements in N rows. If it’s not what it´s desired may be you can first put a Text Aggregator and just after the Iterator to gather all the text and after that insert it.
Instead of using a Update row module use a Bulk update (which accepts an array as input), it will process all the array elements, but will create, also, N rows.
