How should a CSV import handle blank values when updating existing records?

Question

Imagine importing a CSV to update existing SaaS records. A blank cell might mean leave the existing value alone, clear it, or unknown. Treating all blanks the same way could erase useful information or make it impossible to remove outdated values.

What import options would you offer so the behavior is predictable? Should clearing a value require a separate marker or a field-level setting? I am interested in how a preview screen could make the effect clear before the user commits a large update.

hgray · Answer

Make the default explicit and provide a separate way to request clearing a value. A preview should show before and after values, with counts for unchanged, updated, cleared, and invalid fields. Validate required fields before applying changes, and let the user download the rejected rows with reasons so a partial failure is easy to repair.