As of 2.6.7, I'm getting some odd behavior on required fields within repeats. Specifically:
1. When select_one fields used in the last record are selected, the answer shows with a strikethrough
(but if the field was not used in the immediately preceding record (due to relevant setting), it looks normal)
2. Required integer fields are defaulting to zero - with no default value set otherwise
(this is a bigger deal as zero can be meaningful as a value and a user may forget to correctly fill it in)
All data appear to be preserved and do not have strikethrough formatting if repeat records are reviewed (before submitting or from outbox).
This may be related to another reported issue:
select_one choice strikethrough when opened by inbox
Solved! Go to Solution.
Okay, I isolated the bug:
If within a repeat block and using a relevant expression and required = yes, select_one questions will appear with a strikethrough if using a minimal appearance. Likewise integer fields will be pre-populated with zeros. This has occurred with both selected() and string-length() relevant expressions. Having potentially false zeros in the survey is a rather unfortunate bug.
See attached example.
FYI: This occurred on a survey that had been working without error in previous versions of Survey123.
If it helps, I get this if I put a default like 'N' in a yes no list option which is an incorrect entry
maybe points you in a direction ,,
Wow, I didn't expect that to work on an integer field.
I changed the default to 'n' and it shows up (in red) as the default in Survey123 Connect anyway. Haven't tested yet with publishing...
I was able to publish a survey with "number" as a default value for the integer field, which should prevent erroneous zeros in repeats. Thanks for the tip, Mark.
Okay, I isolated the bug:
If within a repeat block and using a relevant expression and required = yes, select_one questions will appear with a strikethrough if using a minimal appearance. Likewise integer fields will be pre-populated with zeros. This has occurred with both selected() and string-length() relevant expressions. Having potentially false zeros in the survey is a rather unfortunate bug.
See attached example.
FYI: This occurred on a survey that had been working without error in previous versions of Survey123.
Using a constraint of '.>0' works as well for preventing zero values in one of my surveys.
Hi Braden,
Thank you for the information.
I wanted to let you know that I am actively investigating this behavior and will update after testing.
Brandon
Braden Burkholder Thank you for bringing these issues to our attention.
I have gone ahead and added the first issue into our internal backlog - strikethrough appearance is seen with select_one questions that are within a repeat, required, and containing a relevant statement in records after the first.
Note: This should not affect the ability to submit surveys nor the values that are submitted.
As for the second issue, where integer fields within a repeat that have relevant statements autopopulate a 0 on records after the first, this should be resolved with the next release of Survey123 (2.7).
You can access the beta through the Early Adopter Community right now if you have an immediate need for the fix. Also, you can checkout some of the other cool features coming with future releases of Survey123!
Thanks again for sharing your discoveries.
Cheers,
Brandon
Thanks for the update Brandon!