Modifying existing Schema (Advanced)
Schema editing should be approached with care, especially once data has already been entered.
This article explains how to modify an existing Schema within a Data Table, and outlines which changes are safe — and which may result in data loss.
Schema editing should be approached with care, especially once data has already been entered.
Where to edit a Schema
You can access Schema editing via:
Data Management → Self-service Data Collection →
Select a Dataset → Select a Dataset Template → Select a Data Table
Important note on data loss
⚠️ Modifying a Schema after data has been entered can result in data loss if the change affects existing data points.
Data loss occurs when previously entered values no longer match the updated Schema structure.
Unsafe Schema changes (will result in data loss)
The following changes are considered unsafe and will cause loss of existing data in affected areas:
-
Adding, removing, or renaming Variables
Previously entered data points will no longer align with the modified Variables. -
Removing or renaming Categories
Data entered against those Categories will no longer match the Schema and will be lost. -
Renaming the Data Table
These actions change how data is structurally referenced, making existing values incompatible.
Safe Schema changes (no data loss)
The following changes are safe and will not result in any data loss:
-
Changing the order of Variables or Categories
-
Moving Variables between rows and columns
-
Adding new Categories to an existing Variable
-
Adding a totals configuration to a Variable that was previously set to None
These updates adjust presentation or extend the Schema without breaking existing data relationships.
Best practice guidance
-
Finalise your Schema design before opening data entry wherever possible.
-
If data has already been entered, review whether a change is safe before applying it.
-
For structural changes, consider duplicating the Data Table or Dataset Template first.
Why this matters
A stable Schema protects data quality, reduces rework, and ensures collected data remains reliable for aggregation and reporting. Understanding which changes are safe helps moderators evolve datasets without unintentionally losing valuable information.
Unsafe operations that may will result in data loss include:
- Adding/removing/renaming Variables will result in data loss as none of the entered data points for the modified Variables will match after this is done.
- Removing/renaming Categories will also result in data loss for the same reason.
Safe operations on schema that will not result in any data loss are as follows:
- changing the order of Variables or Categories
- moving Variables between rows and columns
- adding new Categories to a Variable
- adding a totals configuration for a variable that was previously set to None