Our survey generates a long random string using the Expression Manager ({join(ceil(time()/floor(rand(1000,999999))), substr("abcdefghijkmnpqrstuvwxyzABCDEFGHJKLMNPQRSTUVWXYZ", -rand(1,48),1), substr("abcdefghijkmnpqrstuvwxyzABCDEFGHJKLMNPQRSTUVWXYZ", -rand(1,48),1),substr("abcdefghijkmnpqrstuvwxyzABCDEFGHJKLMNPQRSTUVWXYZ", -rand(1,48),1), ceil(time()/floor(rand(1000,999999))))}).

The problem is that when this is exported in our results, the expression is re-evaluated, so the exported result disagrees with the value the user sees when taking the survey. The only fix I could find is to have the user cut and past the random string into a short-text field which is then properly stored in the results. Is there a better way to store the value of the expression so that what appears in the results agrees with what the user saw on the screen?

Thanks!

I hope that I understand what you mean, but if use the expression as the default value of a text field, the expression is evaluated just once. The text field gets a random value and that is it.
Frans

I haven't tried using the expression as the default value of the text field. I was simply making the survey respondent cut and paste the random string into the text field.

Originally, I tried using an Equation Question type, since this records the value of the equation and returns it in the results report. Unfortunately, when i do this, the reported value in the results is not the value seen by the survey respondent, it is some *other* random value generated using the same expression.

I tried creating a hidden Equation Question to store the random value. I could then display it in text and display it as the default value in a short-text field. When I to this I get the same random string displayed in running text AND as the default text field AND reported in the results file. (I need to know the value of the random string seen by the survey respondent.) The results file also includes a value for the hidden question. The value for that random string, as reported in the results file, is different from the one reported for that question in the visible question that follows it.

So using the default value of a short text field to report the random string will help me get around this problem! Many thanks.

Glad to hear that this solves your problem. But why so complicated? Why do'nt you enter your EM string directly as the default value of the question? You do'nt need the equation question.
Frans

chakali wrote: Originally, I tried using an Equation Question type, since this records the value of the equation and returns it in the results report. Unfortunately, when i do this, the reported value in the results is not the value seen by the survey respondent, it is some *other* random value generated using the same expression.

Strange ?

If it's real with the last version, can you make a example lss file ?
Did the survey have complete index ?

Denis

I'm using Version 2.05+ Build 140811. The lss file I attach will report a random string on the survey web page that is different from the value if the results from an active survey are reported by exporting a CSV file.

If I put the {TheRandomValue} in a short answer question, then it appears the same as it is reported in the results CSV file. If I do the same using a simple text question (what I tried first), the CSV results disagree with what was shown on the survey.

Yep,

No bug here.

The random value shown to user are the randomValue at start of the page, but when you show the page : randomValue are updated, and again when backup.

You just need to put the randomValue BEFORE showing it. (you can sow it too in last page)

Denis
PS: lss fixed attached

Thanks for the help!