Missing data in XLS and CSV file but available on server

Greetings community,

I create a survey with 318 questions that were all mandatory, however when I check the 1297 submissions from enumerators I see blanks on many questions that should not be there! Whether I select “CSV” file download, “XLS” download or even API get requests to either google sheets or tableau! When I investigate individual submissions and try to edit I find that data is captured correctly? I need urgent help on this! Thanks!

Here are the important details I’ve seen that are needed:

I can provided the XML file, the CSV and XLS data files on request thanks!

Hi @isaactumu
Welcome to the community forum. Could you please share a screenshot with one of the forms that you are editing, and its associated data when you download the XLS? I would like to look at that before I begin additional investigations.


1 Like

Thanks, @stephaneloo,

Here are the screenshots (

, 2, and 3) of the Kobo data for your perusal.

I noticed after looking at the data again that because these questions are repeated across multiple surveys (age and gender, across household, business, and farmer surveys, which was one survey file that only shows a specific survey based on selection made by the person doing the enumeration) the extracted file will only show the correct responses after the downloaded excel files are filtered see screenshot (excel 1 and excel 2).

Is there a way to change how this is done or is that how it should be based on how the survey was set up?

Please advise

@isaactumu, we will have a closer look at your issue and get back to you soon.

Hi @isaactumu
Could you please explain this part properly

Hi @stephanealoo,

Now I created a survey with three surveys (See screenshot 1) with it based on selections “Household”, “Business” and “Farmer” see (screenshot 2)

I realised that because this survey is three in one, the data extracted will also be arranged in this manner. I now realised that this is why the excel extracted data has first section blank, secondary section blank and then third section with the correct data (in this case the Farmer survey was the last of the three hence it will be in columns further along in the downloaded excel sheet) because the questions are arranged according to selections made as illustrated in the previous images.

I think I will have to manually clean the data to pick out the correct responses per survey type.

I hope this is clearer. My question, therefore, is it that it is not necessarily that these are blanks or null fields, but it is because of how the survey was created? I expected them to be in a single column for each of the repeated questions regardless of how the survey was created?

Please advise?

This is related to how you created the survey. questions that were skipped and as such they will remain blank.

The repeat groups are different and as such you will get different information. In the future when you design questions that are similar regardless of the survey type, do not place the questions in separate repeat groups if you are interested in having them feed similar columns.


1 Like

Greetings @stephanealoo,

Thanks for the feedback.

Could you advise how I can go about using questions that are similar?

I am not sure how to use them except place them in separate groups depending on the survey to avoid incorrect data?

Hi @isaactumu
This is a structural issue with how you shaped your questionnaire. Could you share the original questionnaire?


1 Like

Greetings @stephanealoo,

Apologies for the late response.

Find attached the original survey (a lot changed to it but that is the general structure). Baseline Survey Questionnaire_6 Oct.docx - Google Docs

Kind regards,

Bonjour isaac, concernant ta question j’ai peut etre diagnotisque le probleme comme ceci.

Dans ton formulaire il y a plusieurs groupes, alors lorsque begin_groupe n’est pas obligatoire (required=yes) les enqueteurs auront tendances a sauter ces questions et les soumissions auront toujours lieu.

Donc, essaye de rendre obligatoire le type begin_groupe et tu ne loupera aucune donnee a mon avis.



Thanks @caredrc_goma for your suggestion

1 Like