I have a survey with a numeric input type question (Q42) followed by an Array (1-5) -- Q43.
The array question (Q43) had a condition on it to only display if :
((Q42.NAOK > " "))
When testing this survey, if I left Q42 blank, then Q43 was hidden, and if I entered a value, Q43 was shown, as expected.
I understand in restrospect that I should have used !is_empty(Q42) instead ... however my testing showed that Q43 was displayed.
I activated the survey, and after collecting data from 10,000+ respondents, discovered that in all cases responses Q43 was blank. Q43 wasn't set to mandatory, but with the number of respondents, it seems unlikely that anyone who entered a value for Q42 would skip Q43.
Can anyone shed some light on what could have happened here? Does LS do any server side validation of EM rules on top of the client-side validation?
I am running LS version 2.73.0 b171219
I've attached a copy of the relevant segment of the survey.
blocka wrote: Does LS do any server side validation of EM rules on top of the client-side validation?
EM validation is client-side (javascript) and server-side (php). There was a time, where everything should be working even without javascript. That aim is no longer valid.
Depending on the questiontopic and the respondents it still is possible, that nobody entered a answer.
I never used > " " in a survey. So there might be a issue/bug with that EM part. This bug can be in the client or the server validation part. The issue can be non-existing in ealier or later versions than yours. So might be hard to reproduce.
One added information:
If you run your urvey "question by question" you see that q43 is NEVER displayed (and so -> no answers)
"group by group" q43 is displayed, but in my opinion that's a bug: Displayed but handled as if not displayed -> no answers.
As you know now, your relevance was not the very best.
Joffm
Volunteers are not paid.
Not because they are worthless, but because they are priceless
Thanks Joffm -- I see that you are right about the question by question -- very odd! Our survey was set to run Group by Group -- do you think we should open a ticket for the issue wrt to question by question?
