我可以使用 <-> 找到 PostgreSQL 中数组之间的距离吗?
Can I find distances between arrays in PostgreSQL using <->?
据我从 this 文章中了解到,在处理几何数据类型时,您可以使用 <-> 距离运算符找到最近的邻居:
SELECT name, location --location is point
FROM geonames
ORDER BY location <-> '(29.9691,-95.6972)'
LIMIT 5;
您还可以使用 SP-GiST 索引进行一些优化:
CREATE INDEX idx_spgist_geonames_location ON geonames USING spgist(location);
但是我在文档中找不到任何关于对数组使用 <-> 运算符的信息。例如,如果我使用 double precision[]
而不是 point
来执行相同的查询,那行得通吗?
显然,我们不能。例如,我有一个简单的 table:
CREATE TABLE test (
id SERIAL PRIMARY KEY,
loc double precision[]
);
我想从中查询文档,按距离排序,
SELECT loc FROM test ORDER BY loc <-> ARRAY[0, 0, 0, 0]::double precision[];
没用:
Query Error: error: operator does not exist: double precision[] <-> double precision[]
Documentation has no mention of <-> for arrays as well. I found a workaround in accepted answer for question, but it poses some limitations, especially on array's length. Although there is an article(用俄语编写),它提出了数组大小限制的解决方法。创建样本 table:
import postgresql
def setup_db():
db = postgresql.open('pq://user:pass@localhost:5434/db')
db.execute("create extension if not exists cube;")
db.execute("drop table if exists vectors")
db.execute("create table vectors (id serial, file varchar, vec_low cube, vec_high cube);")
db.execute("create index vectors_vec_idx on vectors (vec_low, vec_high);")
元素插入:
query = "INSERT INTO vectors (file, vec_low, vec_high) VALUES ('{}', CUBE(array[{}]), CUBE(array[{}]))".format(
file_name,
','.join(str(s) for s in encodings[0][0:64]),
','.join(str(s) for s in encodings[0][64:128]),
)
db.execute(query)
元素查询:
import time
import postgresql
import random
db = postgresql.open('pq://user:pass@localhost:5434/db')
for i in range(100):
t = time.time()
encodings = [random.random() for i in range(128)]
threshold = 0.6
query = "SELECT file FROM vectors WHERE sqrt(power(CUBE(array[{}]) <-> vec_low, 2) + power(CUBE(array[{}]) <-> vec_high, 2)) <= {} ".format(
','.join(str(s) for s in encodings[0:64]),
','.join(str(s) for s in encodings[64:128]),
threshold,
) + \
"ORDER BY sqrt(power(CUBE(array[{}]) <-> vec_low, 2) + power(CUBE(array[{}]) <-> vec_high, 2)) ASC LIMIT 1".format(
','.join(str(s) for s in encodings[0:64]),
','.join(str(s) for s in encodings[64:128]),
)
print(db.query(query))
print('inset time', time.time() - t, 'ind', i)
据我从 this 文章中了解到,在处理几何数据类型时,您可以使用 <-> 距离运算符找到最近的邻居:
SELECT name, location --location is point
FROM geonames
ORDER BY location <-> '(29.9691,-95.6972)'
LIMIT 5;
您还可以使用 SP-GiST 索引进行一些优化:
CREATE INDEX idx_spgist_geonames_location ON geonames USING spgist(location);
但是我在文档中找不到任何关于对数组使用 <-> 运算符的信息。例如,如果我使用 double precision[]
而不是 point
来执行相同的查询,那行得通吗?
显然,我们不能。例如,我有一个简单的 table:
CREATE TABLE test (
id SERIAL PRIMARY KEY,
loc double precision[]
);
我想从中查询文档,按距离排序,
SELECT loc FROM test ORDER BY loc <-> ARRAY[0, 0, 0, 0]::double precision[];
没用:
Query Error: error: operator does not exist: double precision[] <-> double precision[]
Documentation has no mention of <-> for arrays as well. I found a workaround in accepted answer for
import postgresql
def setup_db():
db = postgresql.open('pq://user:pass@localhost:5434/db')
db.execute("create extension if not exists cube;")
db.execute("drop table if exists vectors")
db.execute("create table vectors (id serial, file varchar, vec_low cube, vec_high cube);")
db.execute("create index vectors_vec_idx on vectors (vec_low, vec_high);")
元素插入:
query = "INSERT INTO vectors (file, vec_low, vec_high) VALUES ('{}', CUBE(array[{}]), CUBE(array[{}]))".format(
file_name,
','.join(str(s) for s in encodings[0][0:64]),
','.join(str(s) for s in encodings[0][64:128]),
)
db.execute(query)
元素查询:
import time
import postgresql
import random
db = postgresql.open('pq://user:pass@localhost:5434/db')
for i in range(100):
t = time.time()
encodings = [random.random() for i in range(128)]
threshold = 0.6
query = "SELECT file FROM vectors WHERE sqrt(power(CUBE(array[{}]) <-> vec_low, 2) + power(CUBE(array[{}]) <-> vec_high, 2)) <= {} ".format(
','.join(str(s) for s in encodings[0:64]),
','.join(str(s) for s in encodings[64:128]),
threshold,
) + \
"ORDER BY sqrt(power(CUBE(array[{}]) <-> vec_low, 2) + power(CUBE(array[{}]) <-> vec_high, 2)) ASC LIMIT 1".format(
','.join(str(s) for s in encodings[0:64]),
','.join(str(s) for s in encodings[64:128]),
)
print(db.query(query))
print('inset time', time.time() - t, 'ind', i)