Scalar filtering optimization #38756
-
|
I saw this quote in the blog.. I don't quite understand what it means, can you tell me more about it. "So if we can figure which buckets to compute, we can just filter vectors in those buckets to greatly reduce computation load. " Is this optimization now implemented？ |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
From v2.4.8, Milvus provides a new feature "clustering key": You can define a clustering key in the collection schema: With this clustering key, data is arranged by the values of the "dummy" field. When you call search(filter="dummy == 8"), the search engine can quickly know which segment contains the data with "dummy == 8", other segments are skipped, to improve the search performance. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you. That's great! |
Beta Was this translation helpful? Give feedback.
From v2.4.8, Milvus provides a new feature "clustering key":
https://milvus.io/docs/clustering-compaction.md#Clustering-Compaction
You can define a clustering key in the collection schema:
With this clustering key, data is arranged by the values of the "dummy" field. When you call search(filter="dummy == 8"), the search engine can quickly know which segment contains the data with "dummy == 8", other segments are skipped, to improve the search performance.