Brian Cassidy (Oct 22 2025 at 13:58): Brian Cassidy (Oct 22 2025 at 13:58):We had a user report that when they uploaded a CSV file, they got an error along the lines of "invalid header row. one of the cells is empty".
The user's file was fine, it was formatted in such a way that it didn't use column names in the first cell for some columns.
Is this a needlessly alarming error, or is the intention that any csv files that aren't pure header-per-column formatted be marked as not "proper" ?
Philip Durbin 🚀 (Oct 22 2025 at 14:43):Sounds similar to this issue:
Tabular data without headers with first row cell content longer than 255 characters cause exception for document ingestion #7386
Philip Durbin 🚀 (Oct 22 2025 at 14:44):We do expose variable names in certain export formats like DDI and Croissant. They come from the first row, the header.
Philip Durbin 🚀 (Oct 22 2025 at 14:45):Can you provide us with a CSV file that triggers the error?
Brian Cassidy (Oct 22 2025 at 16:49):Sent to your gmail. All the CSVs upload look similar to that.
Brian Cassidy (Oct 22 2025 at 16:50):i see they have a kind of annotation as the first line by itself, but even the real data part wouldn't have column headers for the first two columns. Perhaps CSV "etiquitte" needs to be learned? :)
Philip Durbin 🚀 (Oct 22 2025 at 17:03):Yeah, wouldn't hurt, I'd say. There's a header! But it's on line 2. :smile:
Brian Cassidy (Oct 22 2025 at 17:05):Would it still throw the error because of those first two columns?
Philip Durbin 🚀 (Oct 22 2025 at 17:26):probably
Philip Durbin 🚀 (Oct 22 2025 at 18:18):try it and see, I guess :smile:
Last updated: Oct 30 2025 at 06:21 UTC