Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
fix: filter the vec index in function Index::scalar_index_info (#3000)
Create the vector index error when the schema has nullable field. ![image](https://github.com/user-attachments/assets/d323b9e0-6d39-4293-92b7-3ba3adfc8b2b) When the vector column had build index already, the dataset scaner will assign the filter in function `compute_partitions`. ```python if dataset.schema.field(column).nullable and filter_nan: filt = f"{column} is not null" else: filt = None ``` The scanner will use the scalar index scanner and this scanner treat the vector index as scalar index(btree index). This pr will filter the vector index in `scalar_index_info` and it will only return the scalar index now.
- Loading branch information