<aside> ❗ Databricks Runtime version 11+ is necessary for this operation.
</aside>
Let’s assume we have a feature store table odap_feature_store.client_features
like this
client_id | timestamp | f1 | f2 | f3 |
---|---|---|---|---|
1 | 2020-01-01 | 100 | Praha | null |
2 | 2020-01-01 | 200 | Brno | null |
3 | 2020-01-01 | 300 | Bratislava | fjwoiejfw |
1 | 2020-01-02 | 150 | Praha | null |
2 | 2020-01-02 | 220 | Ostrava | null |
3 | 2020-01-02 | 280 | Brno | null |
and a metadata table odap_feature_store.client_metadata
like this
feature | description | other columns |
---|---|---|
f1 | Feature 1 | … |
f2 | Feature 2 | … |
f3 | Feature 3 | ... |
To delete a feature f3
you need to do two things:
ALTER TABLE odap_feature_store.client_features DROP COLUMN f3
Result
client_id | timestamp | f1 | f2 |
---|---|---|---|
1 | 2020-01-01 | 100 | Praha |
2 | 2020-01-01 | 200 | Brno |
3 | 2020-01-01 | 300 | Bratislava |
1 | 2020-01-02 | 150 | Praha |
2 | 2020-01-02 | 220 | Ostrava |
3 | 2020-01-02 | 280 | Brno |
DELETE FROM odap_feature_store.client_metadata WHERE feature = "f3"
Result
feature | description | other columns |
---|---|---|
f1 | Feature 1 | … |
f2 | Feature 2 | … |