Why are there ObjectID gaps in Survey123 data?

251
4
09-29-2020 05:32 PM
AnnekaImkamp
New Contributor

Hi - I have a Survey123 that has been public for over a month and received over 5,000 submissions. Until today, the ObjectIDs were consistent and in order (i.e. 1 to 5851). However, there was a 40-minute time period this morning during which 14 OIDs got skipped (never existed in the data), with a few that came through in between (we are missing OIDs 5852-5857, 5859-5863, and 5870-5872). Is there an explanation for this? Without having any jumps in the OID range until now it seems concerning, as if somehow those entries were lost or not properly transmitted through the AGOL servers. If jumps in OIDs are to be expected, why would this only happen now after a month of consistency? Since the issue this morning, the OIDs have been populating in order, as before.

Thank you!

0 Kudos
4 Replies
Philip-Wilson
Esri Notable Contributor

Hi Anneka,

Do you know exactly what date and time the issue occurred? If so, please contact Esri Support and they will open a case to investigate if there was any interruptions or down-time during that exact time and if it affected your ArcGIS Online organization.

It is not expected there would be any gaps in ObjectID unless the data has been deleted after it was successfully submitted, or there was an issue with the server at the time.

Regards,

Phil.

0 Kudos
FayGlunt
New Contributor II

Hi, Phil.

We are skipping object IDs too.  Lots of skips in December 2021.  Should we open a case also?

Thanks, Fay

0 Kudos
FayGlunt
New Contributor II

One more question, Phil.  :)

If the submission is initiated, does the object ID get reserved for that submission and the sequential number range incremented?  But, if the submission is interrupted, it is rolled back and not committed to the database?  Thus creating a skipped object ID?

0 Kudos
Philip-Wilson
Esri Notable Contributor

Hi @FayGlunt,

The objectid is created when the record is added to the feature layer (database), based on the previous existing objectid and incremented by 1. If the transaction is rolled back, there is no objectid created (as no record), so the last known objectid is still the same. So when the next record gets added it should get the next sequential objectid, there would be no gap.

The only time this may not occur is if multiple records are being created and deleted at the same time by different users on different devices. In that case gaps may be created if multiple records get created at once, and some then deleted but newer records were created before previous records deleted. That's the only way I can think this would happen.

Regards,

Phil.

0 Kudos