VECTOR_COSINE_SIMILARITY

Snowflake

Menghitung cosine similarity antara dua vector untuk semantic search.

Tipe hasil: FLOATDiperbarui: 6 Jan 2026

Syntax

SQL

VECTOR_COSINE_SIMILARITY(vector1, vector2)

Parameter

vector1VECTORwajib

Vector pertama

vector2VECTORwajib

Vector kedua

Contoh Penggunaan

Find Similar Documents

SQL

1 SELECT
2   d2.title,
3   VECTOR_COSINE_SIMILARITY(d1.embedding, d2.embedding) as similarity
4 FROM documents d1
5 JOIN documents d2 ON d1.id != d2.id
6 WHERE d1.title = 'Machine Learning Basics'
7 ORDER BY similarity DESC
8 LIMIT 5;

Cari dokumen yang mirip.

Kembali ke Snowflake Functions

1	SELECT
2	d2.title,
3	VECTOR_COSINE_SIMILARITY(d1.embedding, d2.embedding) as similarity
4	FROM documents d1
5	JOIN documents d2 ON d1.id != d2.id
6	WHERE d1.title = 'Machine Learning Basics'
7	ORDER BY similarity DESC
8	LIMIT 5;