Missing part of surveys data

Dear ODK experts,

I'm currently implementing an endline data collection. After the second week of the collection when I've tried to export the data using briefcase, I got a dataset without all surveys sent in the first week; this is the first time for me to experience a such situation. Fortunately for me I save the first week dataset somewhere else. But I would like to get all the data in one CSV from ODK to allow easy manipulation or treatment.

I'm using ODK collect for the survey and ODK briefcase v1.17.3 on Mac
I've tried many times to clear pull story but still have the same problem.

Could you please explain to me what could be the problem? Do you have any idea to solve the situation? Any suggestion will be welcome. Thanks in advance.

Hi @gdodo,

Sorry for the bad experience that you have come across.

Since you have access to the device, are you sure while moving the data files you didnt delete the data files on the device? ODK saves the data in the xml files saved in the odk folder. This data cant be deleted even when uninstalling the app accidentally.

There might be possibility of deletion on some step which requires investigating.

Is this data already in CSV format, or the raw files from the device?

If the data is in CSV format, then you have to append manually the two dataset and you can find the duplicates using the KEY column. If you have the raw files from the device I think you can just combine the submissions of the same form, then export.

Since you are doing offline collection, I would advise keeping copies of the ODK folder from each device as a backup before/after exporting.

Hello @dicksonsamwel,

First of all thank you very much for your help.

The data I saved is a dta format (stata); I got this extension when I run the odkmeta do file. The csv format has been overwritten when exporting data from odk briefcase

I think this is the solution. I will try to convert the .dta format in csv and then append it manually. For the information I have also try to append the two datasets using Stata but there was some mismatch in the dataset at some point.

Thanks for your advice

1 Like