Dataverse Tip #9 9 Oct 2026

Your table has twelve columns you did not create, and one of them rewrites Created On

Every Dataverse table carries twelve columns nobody created - and mapping a source date to overriddencreatedon puts it in createdon instead.

Every Dataverse table you build comes with twelve columns you did not create. They are createdby, createdon, createdonbehalfby, modifiedby, modifiedon, modifiedonbehalfby, importsequencenumber, overriddencreatedon, statecode, statuscode, timezoneruleversionnumber and utcconversiontimezonecode. A table that uses owners or business process flows carries eight more, and ownerid, owninguser, owningteam, owningbusinessunit, processid, stageid and traversedpath are the ones you will meet.

overriddencreatedon is the one people get wrong. It does not hold the date its name suggests. Map your source system’s created-on column to it during an import and Dataverse writes that value into createdon and puts the import time into overriddencreatedon. Map nothing and createdon becomes the day you ran the import while overriddencreatedon stays empty.

So after a migration, a report that reads overriddencreatedon as the date a record was really created is reading your migration date, and a report that reads createdon is right, even though the usual assumption is that a migration flattened it. The override also needs the prvOverrideCreatedOnCreatedBy privilege. That is usually why one team’s dates survive a migration and another team’s do not.

importsequencenumber does the same job for imports. Every import job stamps one sequence number on every record it creates, so a single number finds all the rows from one import when something goes wrong halfway through.

utcconversiontimezonecode and timezoneruleversionnumber are the platform’s own time zone bookkeeping and sit idle on most tables.

This matters most in two places: a migration that has to keep the original dates, and a report where “created on” is supposed to mean something. If you are about to build one of these twelve columns yourself, you almost certainly do not need to.