A NOTE has been added to this issue.

I am sorry but this survey is too big to test the issue.
Please attach a survey with a minimum test case (max. 2-3 questions) that reproduces the error.

Attached just the first 4 questions it still does it in export Excel - talk to Sam Mousa he asked me to raise it.

I have left the rest of the algorithms in.

I am sorry but I don't see what's wrong here. The bug report is very confusing. Please state the issue and how to reproduce it using the attached survey. Thank you.

Have you contacted Sam Mousa as suggested. sammousa@gmail.com

1. the algorithms for jc1t to jc4t work fine when on screen and are un-hidden the results show.
2. When you export as an excel or csv file only jct1 returns a result - jc2t to jc4t are blank returning nothing.

Hope this clear - the problem is in export.

If I import the attached survey here it is full of EM errors, mostly referring question codes are wrong. The survey is still way too long and pretty much unusable for bug testing.

Anyway, it looks like you need to set the output of the equation to numeric (advanced question attributes) otherwise the result is a string and you are comparing to =1 not ='1' so it fails.

Anyway, I don't think it is a bug.

Disagree

here are the screen shots - does not make sense because why does jc1t work and jv2t jc3t and jc4t do not work.

I know the other return codes because I did not take them the them out for this test sample.

Attached is screen shot of all in one you can see the answers are correct. Also attached is excel spreadsheet look at col CQ CR CS and CT - CQ returns the correct answer the others are blank.

Hope this helps the problems appears to be a bug in sending to excel or csv file - Sam reckons there is an added space?

have you talked to Sam? see attached screen shots this appears to be an excel cvs problem within LS?

I will try to clear some things up.

A strong indicator that it is in fact a bug is the fact that different behavior is observed with client side evaluation (survey all in one, javascript generated by EM) and server side evaluation (survey question by question).

In the second case, server side, a space is added in front of the answer code.

Further, the problem is not in export but simply in evaluation of the value of a variable. The strange thing is that this does not happen for all questions of that question type. I have also confirmed that the question codes are stored correctly (also in the database).

I do not directly know what is causing this issue, except that it is something EM is doing wrong.

I cannot reproduce this here as I still have no testing case to reproduce. The attached survey is way too big and full of errors. This already has taken too much of my time trying to find out what the issue is in the first place. I am unassigning this for now. Maybe someone else likes to take this one.

There are space before expression in lss file for jc3 and jc4, not for jc1.

Use trim()
or try with Équation settings to numeric only .

Not a bug here, EM result can be 'only space' if needed.

PS: space is visible in editor in txt(upload screen)

Denis

Thank you

I have applied the fix which is really a work around because it begs the question why this does not happen in jc1. It only seems to happen with certain algorithms?

Anyway this is what I have used for jc2 etc

{if(Q01_sq1 =="1",trim(1),"")}{if(Q01_sq1 =="5",trim(5),"")}

At least I have a fix

Thanks again

It's not in algorithm, it's in question text.

To use equation, best is to allways deactivate HTML editor (maybe a plugin idea ? ;) ).
Here, i think it's best if you use :
{if(trim(jc4) == "1", "A Future","")}{if(trim(jc4) == "5", "Not A Future","")}
Denis

Space in equation question text don't need to be removed.

Add some more information

With more testing : withlast version : seems to happen only with non-breaking space.

Still an issue with 'only numeric" and & nbsp ;5 : i think we have to remove & nbsp ;   if Advanced settings is set ? (or not ?)

I think of 2 way to fix it (and only for advanced setting/numeric)
floatval(question.text) or
(is_numeric(trim(question.text))? trim(question.text) : "";

Carsten : your advice ?
(see LSA file EQ1 in DB are  5: it's NOT numeric)

Just a point the reason I put in the question algorithm is that long before the text jc4t was even added jc4 was always adding a hatted A before the number in the csv file.

I raised this with Sam last year and he thought it was because I was opening up the csv file in excel - this has proved not to be the case.

So in short jc1 was always OK but jc2, 3 and 4 has always returned the hatted A before the number in a csv file.

The current fix I am using returns the text to the SVG file when called from LS in our app. So the hatted A in the question algorithm appears to have been the cause.

Hope this adds something.

Denis, feel free to add it.

No change required.

Put another bug for equation/advanced setting/numeric