COSINE_SIMILARITY

Applies to: ✅ Data 360 SQL ✅ Tableau Hyper API

Computes the cosine similarity between two vectors, measuring directional alignment.

1COSINE_SIMILARITY(<array1>, <array2>)

• <array1>: The first vector (array(real) or array(double-precision)).
• <array2>: The second vector (array(real) or array(double-precision)).

Returns a double-precision value between -1 and 1, where 1 indicates identical direction, 0 indicates orthogonal, and -1 indicates opposite direction.

• Both input arrays must be of the same element type and length.
• If any element is NULL, the result is NULL.
• Zero vectors cause division by zero.
• For optimal performance, use array definitions with non-nullable elements.

Compute the cosine similarity between two vectors.

1SELECT COSINE_SIMILARITY(array[1, 2, 3], array[-1, 2, -3]) AS result;

Returns approximately -0.42857, indicating the vectors are somewhat opposed in direction.

• Array Data Type