HAMMING_SIMILARITY

Applies to: ✅ Data 360 SQL ✅ Tableau Hyper API

Calculates the similarity between two quantized (binary) embeddings using the normalized Hamming distance.

• bytea1: The first binary-quantized vector.
• bytea2: The second binary-quantized vector.

Returns a double-precision value between 0 and 1, where 1 indicates identical binary vectors.

• Both inputs must be bytea values of the same length.
• This function provides fast, approximate similarity. Use exact methods when precision is critical.

Compare two binary-quantized vectors.

Returns 0.75, indicating 3 out of 4 bits match between the two vectors.

• Array Data Type