we had a working form but list of choices of one question became too long so now I am trying to have it as a csv attached to KOBO as media. ODK form validation is ok, the upload is just fine. I can download the form into ODK just fine too. but when opening the form, it doesn't open and instead gives an error message attached below
2. What steps can we take to reproduce this issue?
3. What have you tried to fix the issue?
i have looked at all other questions in the community but i havent seen one like what i have.
I have emailed support too but I was advised its better to ask here.
4. Upload any forms or screenshots you can share publicly below.
The error message is correct, the cause of this error is that you have no name column.
If you look closely at your CSV, you will see that your column is called name not name. remove the trailing space and it will work. Very hard to spot error!
Annoyingly, it WILL work exactly as expected in Enketo, and kind of work in Webforms, but the slightly off CSV causes some strange behaviour.
@gareth a couple for you!
With the name value CSV:
In Webforms selecting any item selects all items with search appearance, I changed the appearance from search to minimal autocomplete and it did the same thing.
Changing to columns appearance let the multiple select sort of work in WF, but the followon questions that had relevance won't appear, and clicking on the label for any option always selects only the first option, not the one clicked, must click in the checkbox only to get the desired selection.
Here I'm moused over Bats, see Aardvark is darker blue. Clicking on Bats then selects Aardvark
Thanks for trying it out and finding this discrepancy.
I'm wondering about what the correct solution here is. We could either trim the column name and have it just work, or display an informative error when the form is loaded. I'm leaning towards the informative error because that would also be helpful when the column name is wrong for other reasons too. This would also be consistent with Collect which is nice because it means errors would be caught in the draft stage before they're in use.
I agree the danger is that if WF silently fixes it like Enketo appears to, then Collect will still error, so behaviour is inconsistent - I initially only tested this form in Enketo / WF so I saw working / strange behaviour and was chasing that before I also tried Collect and realised it was the column name.
so, yeah, if WF returned a similar error about a malformed CSV then it's consistent and catches errors earlier. The only possible improvement I can think of is that the CSVs are validated by Central at upload to catch it even earlier, but perhaps unnecessary / overcomplicated especially with a gradual shift from CSVs to entities.
