In the data exported via API some submissions are missing information in the automatically created columns, i.e. version, uuid, submission_time etc.
It seems to happen randomly, since other submissions submitted by the same device (using KoboCollect and the same version of the form) submitted on the same day do not have this issue.
And even more strangely, when I check the data in KoboToolbox, all the missing information in the api export for the submission concerned is there though.
(screen shot shows the data for the submission in row 125 in the Excel sheet)
Has anyone experienced this before?
Do you have an idea why not all of the data is exported via API even though it exists in KoboToolbox?
Thanks a lot for your help!
Welcome back to the community, @Abraxas! Could you share the API you are using so that the community could add to it?
I am using this link for the API:
Let me know if you need other information as well.
Thanks for your support!
@Abraxas, could you ensure the source of data collection you made for those where some meta data are missing?
Sorry, I don’t understand.
what should I ensure?
What did you use (
Collect android app or
Enketo) to collect data for these missing metadata?
We are using the android App “KoboCollect” to collect the data.
Collect android app for both those present and also for those where the metadata is missing?
But the strange thing is, is that the metadata is only missing in the excel sheet that is connected with the api-export.
If I check the data in KoboToolbox, all metadata is there.
Do you have an idea why the API is behaving like this?
I have just received new data from an interviewer. And while most of the submissions from that interviewer do have the submission time and date in the API exported data, there is again one submission that is missing this data.
I opened the concerned submission in KoboToolbox and resubmitted it but with no success / the submission time and date is still missing in the API-exported data.
Also, I have 27 submissions in the API data-export with random data that do not correspond to the columns at all (and are also nowhere to be found in the KoboToolbox Account);
If you compare the rows loaded in the API data export (3463) with the submissions loaded in the KoboToolbox Account (3436), the difference it 27. So I assume, these are the 27 rows with the random data:
Have you collected the data?
The data was collected by one of our interviewers using the android App KoboCollect.
Hi @Abraxas, the difference in submission count is a sync issue. Can you please send me a private message with the following details so I can sync it from our side:
- server (OCHA/HHI)
Regarding the missing data values, that is more concerning. Let me sync the submissions first and see if the issue persists, otherwise I will investigate further.
In the meantime, I found a work-around for the issue with the missing metadata (submission date) as well as the additional random submissions;
I realized this “bug” only appears when exporting the data in the csv-format. When I used the api-link with the xls-format the exported data was correct and complete, no missing data, no strange additional random submissions.
So, I changed all the data sources to .xls-format and our database was fine again…
Hope, this will be useful for other users as well.
@Abraxas, thank you for sharing the solution with the entire community! This should be very helpful. Expecting your same level of support in the community to make the community a learning and sharing platform.