Dear @Kal_Lam,
Any news on this. We really are waiting for a solution. We even don’t have a workaround: As far as we tried, the problem cannot be solved by relevant marking a cases with Name (or extra variable) as “deleted”, as this will not remove the repeat itself.
We also don’t want to use manual repeat as there are even more reference problems (and a confusing UI).
Thanks in advance
Kind regards
Wolfgang
Dear Stephane,
THANKS. The behaviour should be like in Enketo /Preview(but only without pages styles there):
If the repeat_count is reduced, the records are automatically emptied/removed, not only hidden as with relevant.
You can start with 3, enter data for the 3. Then go back and reduce to 2. Look at the cases/data. Then go back again and reset to 4. The data are already gone.
The worst at the moment in KoBoCollect is, that even finalize and save do not clean up anything. So, also if you have a count after the repeat, it will stay to value before the reduction.
On the opposite, it works like expected if you increase the repeat_count when you go back.
My feeling is it might be a problem of changed-event handling together with repeat_count.
Kind regards
Wolfgang
Hi @wroos
This is definitely an issue with the form; we, therefore, need to discuss this a bit and then get back to you (@Kal_Lam) let us discuss again.
@Kal_Lam@Xiphware@Josh, I’d like to revive this thread, since we are still facing the same problem as described in the OP.
Just today, I had to write an email to one of our data collector partners, explaining the cause of an inconsistency in the dataset that they had collected. I would really appreciate if someone could look into this – the problem is easy enough to reproduce with the instructions above, and data consistency should be a priority for ODK/Kobo, shouldn’t it?
Hi @Sjlver, Kobo is not involved in the development of (Kobo)Collect, only ODK. Can you please open an issue in GitHub repo for this: Issues · getodk/collect · GitHub
