Our team in Indonesia recently added survey question blocks with two small problems:
no English label
long random character strings for data column names
Consequently, I couldn’t work with the data to do analysis. So, I went and added English labels and conventional variable names in the form builder. I downloaded the new survey form as an XLS just to make sure everything was right, and it was.
However, when I redownloaded the data using the XML values and headers, the original variable names were still there as opposed to the new column names I had created. So there is an incongruence between the form and the data.
Maybe this is because the data was originally gathered using the previous column names? or perhaps I need to redeploy the survey for the change to take effect on the data export? Not sure here. Any guidance much appreciated!
When you changed the column names, you essentially created a new question, e.g., let us say you had a column name such as “Jenis_kelamin,” and you changed it to “gender”. The data collected under Jenis kelamin will remain in the column marked Jenis kelamin, and it will not be transferred into the column gender. This is because this is treated as a new question. So as it is, you cannot change the column name midway and have data remapped to it.
Thanks Stephane, I think that makes sense. So even though it is the same question in form builder (I didn’t delete the question and create a new now - only edited the English label and column name of the existing question), it will appear as a new question in the data? Meaning that if I were to redploy the form and new submissions came in, that data would have the new column names, but the old data would have the old column names?
If this is the case, I’ll clarify with this last question: Once surveys have been submitted, the associated data cannot be changed as it appears in its raw exported form by making any changes to the form on the kobo ui, correct?