-
Notifications
You must be signed in to change notification settings - Fork 356
Closed
Milestone
Description
Feature Request / Improvement
When we accumulate changes when doing updates to the metadata, we want to accumulate the changes in the metadata.
For example, if you first add a new column to a table, and then directly write some data, it should take the updated schema. The game goes for when you update the table properties and do a write after that, it should take the table properties into account.
This is an important concept that's currently not working of PyIceberg. For example, an upsert operation in PyIceberg is modeled as first a delete, and then an append operation. This way we want to stack the operations.
HonahX
Metadata
Metadata
Assignees
Labels
No labels