Warning
SingleStore 9.0 gives you the opportunity to preview, evaluate, and provide feedback on new and upcoming features prior to their general availability. In the interim, SingleStore 8.9 is recommended for production workloads, which can later be upgraded to SingleStore 9.0.
Using Vector Functions
On this page
SingleStore provides a native vector data type and supports similarity searches over vector data using k-nearest neighbor search (kNN) and indexed Approximate Nearest Neighbor (ANN) search.
Resources
The following resources provide information about SingleStore's vector processing:
-
Vector Type: The native vector data type.
-
Working with Vector Data: Vector similarity search.
-
How to Bulk Load Vectors: Load larger vector data sets using a binary format.
-
Vector Indexing: Indexed Approximate Nearest Neighbor (ANN) vector similarity search.
-
Hybrid Search - Re-ranking and Blending Searches: Combine indexed (ANN) vector similarity search and full-text search.
-
Tuning Vector Indexes and Queries: Adjust the index building and searching parameters to meet your application's needs for query accuracy and performance.
The following functions are available for the vector type:
-
DOT_
PRODUCT: Calculate the dot product of two vectors. -
EUCLIDEAN_
DISTANCE: Calculate the Euclidean distance between two vectors. -
Cosine Similarity and Cosine Distance: Calculate the cosine similarity and distance metrics.
-
SCALAR_
VECTOR_ MUL: Multiply a vector by a scalar value. -
VECTOR_
ADD: Add two vectors. -
VECTOR_
ELEMENTS_ SUM: Sum all the elements in a vector. -
VECTOR_
KTH_ ELEMENT: Find the kth element of a vector. -
VECTOR_
MUL: Multiply two vectors element-wise. -
VECTOR_
NUM_ ELEMENTS: Count the number of elements in a vector. -
VECTOR_
SORT: Sort a vector. -
VECTOR_
SUB: Subtract one vector from a second vector.
Last modified: April 30, 2024