Removing schema field - handling existing document data

Good question! When you remove a field from your Sanity schema, the existing data in documents that have that field will not be automatically deleted. This is actually a feature designed to prevent unintended data loss.

What happens to the existing data?

The data remains in your documents in the Content Lake. Sanity treats schema changes as non-destructive by default - changing your schema doesn't cascade changes to existing content. This helps prevent breakage for applications that might still rely on that data.

What you should consider doing:

1. Mark the field as deprecated first (optional but recommended for production projects)

   You can use the deprecated property in your schema to mark fields that shouldn't be used anymore:

   defineField({
     name: 'oldFieldName',
     type: 'string',
     deprecated: {
       reason: 'This field is no longer used'
     },
     readOnly: true // prevents further edits
   })

2. Decide if you need to clean up the data

   You have a few options:

   • Leave the data as-is (it won't hurt anything)
   • Use a content migration to remove the field data from existing documents
   • Use the sanity documents validate CLI command to check which documents still have that field

3. Run a migration to remove the data (if needed)

   If you want to actually remove the field data from existing documents, you can create a migration script using sanity migration create and use the unset() helper function.

Where to find more information:

The main documentation page you want is: Migrating your schema and content

Also helpful:

• Content migration cheat sheet - copy-paste examples for common migration tasks
• Schema migration principles - best practices for handling schema changes

The short version: you're safe to remove the field from your schema without worrying about data loss. The old data will just sit there harmlessly unless you explicitly remove it with a migration.