Hi @hphaoag, I am very sorry for the trouble.
Do you also run into this problem when fetching only a partial sheet on each workflow execution? You can specify a range using the Data Location on Sheet option on the node:
Perhaps you can give this a go against your actual dataset?
Second step
If a smaller range is working fine, you could consider fetching only chunks of your sheet with each workflow execution. This is a bit tricky as n8n will automatically assume the first row contains your column headers/field names. I believe the below workflow should handle the conversion, but let’s see if the smaller range works in the first place.
This workflow would transform this dataset:
Into this one:
