I use the lastest build of limesurvey (Version 2.05+ Build 150520)
I have problems with export of text fields to spss. I have a simple short text field named "arts" (docter in duch). Here participants fill in the name of their docter. When I export this field to spss, the export module generates the following syntax:

GET DATA
/TYPE=TXT.
.
.
/VARIABLES=
.
.
V26 F1
.

RENAME VARIABLE ( V26 = Arts ).

so arts incorrectly becomes a one-digit numerical variable. This could be correct if the string values would correspond with var labels, but this is not the case. Is this an error? This does not always happen. I had the impression that this happens if the first case is empty, but that is not so. In some cases the incorrect datatype is not F1 but A1

What I do as a workaround is that I change "v26 F1" into "v26 A16".
Then everything works fine.

Does others have the same problem? Is there a real solution for this?

We also encountered problems in SPSS output. Not only field codes are wrongly coded, but also when you use extra fields in the token table, the SPSS goes completely out of hand.

We solved this by adapting the export_helper.php file and submitted this for being taken up in LimeSurvey core. Until now this has not happened, I guess I still have to describe what we have done.

Just try is this helps you. The unzipped file should go into /application/helpers

Thanks Tammo.
Your solution works very well.
Frans Marcelissen, DigiPsy, Leende

Thank you for testing it. Which version of LS do you use?

I tested it on Version 2.05+ Build 150520. This is my testing site. I will install and test it om my production site later. I hope it soon will be integrated in the core source.

github.com/LimeSurvey/LimeSurvey/pull/296#issuecomment-89249966

I asked mfaber to assist me, because I feel like being in a deadlock: I have to test something I have never used...

I use limesurvey with r and spss every day. If you need someone to test these kind of things, pleas contact me.

Well, a first check could be: does anything go wrong when exporting to R when you use this new export_helper.php?

But mfaber asked to test with STATA, do you use that also?

I do'nt use stata. exporting to R works fine.
But to be honest: I normally don't use this way of importing into R, I alway read the data of limesurvey directly into r using mysql or rpc2.

OK, thanks. I'll ask around at the forum...

Tammo

Is that a bug only related to question type short text? Or is the complete spss export wrong?
Seems to be a bug related to Build 150520.
2.06 has been marked stable. Not sure if a fix will be redistributed for 2.05.

What is the bugtracker id for this bug report?

We did not enter a bug report, we just offered a fix via Github. We will check if the bug still exists.

BTW: it was first encountered in the 150211 build, but it may have existed before that.

My 141210 build has the same problem. I did not notice it before.
The fix of tammo appears to work fine. I have not problems waiting for the 2.06 release for the official fix.
Frans

fransmarcelissen wrote: My 141210 build has the same problem. I did not notice it before.
The fix of tammo appears to work fine. I have not problems waiting for the 2.06 release for the official fix.
Frans

Did you test STATA as well?

See www.limesurvey.org/en/forum/installation...h-an-export-to-stata

I did not test STATA. And I would nog say that I fully tested spss and r: I just used it and noticed that in my case everything works fine. I would nog call that a full test.