Error: could not convert string to float when executing SELECT or ADD query with cosine_distance #114
Comments
SharkSyl
changed the title
|
I changed my code from asynchronous to synchronous, and it works normally. Why is this? Does it temporarily not support this kind of asynchronous operation? Dependency versions： name : sqlalchemy
version : 2.0.37
description : Database Abstraction Library
dependencies
- greenlet !=0.4.17
- typing-extensions >=4.6.0
required by
- alembic requires >=1.3.0
name : asyncpg
version : 0.30.0
description : An asyncio PostgreSQL driver
The entire code is as follows： # -*- coding:utf-8 -*-
from sqlalchemy import URL, PrimaryKeyConstraint, Index, BigInteger, TIMESTAMP, func, select, insert, text, event
from sqlalchemy.orm import Mapped, mapped_column
from sqlalchemy import create_engine
from sqlalchemy.orm import declarative_base, sessionmaker
from pgvector.sqlalchemy.vector import VECTOR
from pgvector.utils.vector import Vector as VectorUtils
url = URL.create(
drivername='postgresql',
username='postgres',
password='vector123',
host='localhost',
port=5433,
database='vector_demo',
)
engine_project = create_engine(url, echo=True, future=True, pool_pre_ping=True)
session_target = sessionmaker(bind=engine_project, autoflush=False, expire_on_commit=False)
@event.listens_for(engine_project, "connect")
def connect(dbapi_connection, connection_record):
from pgvector.psycopg2 import register_vector
print('Init PGVector')
register_vector(dbapi_connection, globally=False, arrays=True)
Base = declarative_base()
class ORMEmbeddingTest(Base):
table_name = "orm_embedding_3"
__tablename__ = table_name
__table_args__ = (
PrimaryKeyConstraint('id', name=f'{table_name}_pkey'),
Index(
f'{table_name}_embedding_index',
'embedding',
postgresql_using='hnsw',
postgresql_with={'m': 16, 'ef_construction': 64},
postgresql_ops={'embedding': 'vector_cosine_ops'}
)
)
id: Mapped[int] = mapped_column(BigInteger, primary_key=True, autoincrement=True)
embedding: Mapped[list] = mapped_column(VECTOR(3), nullable=False)
created: Mapped[str] = mapped_column(TIMESTAMP, server_default=func.now(), nullable=False)
modified: Mapped[str] = mapped_column(TIMESTAMP, server_default=func.now(), onupdate=func.now(), nullable=False)
def create_table():
"""创建数据库表"""
Base.metadata.create_all(
engine_project,
tables=[Base.metadata.tables[ORMEmbeddingTest.__tablename__]]
)
def drop_table():
"""创建数据库表"""
Base.metadata.drop_all(
engine_project,
tables=[Base.metadata.tables[ORMEmbeddingTest.__tablename__]]
)
def run_test():
drop_table()
create_table()
query_embedding = [1, 3, 4]
db_obj = ORMEmbeddingTest(
id=99,
embedding=query_embedding
)
with session_target() as db:
db.add(db_obj)
db.commit()
try:
with session_target() as db:
select_test = db.execute(
select(
ORMEmbeddingTest.id,
ORMEmbeddingTest.embedding.cosine_distance(query_embedding).label('score')
).order_by(text('score'))
)
records = select_test.fetchall()
for record in records:
print(record.id, record.score)
except Exception as e:
print(f'orm insert:{e}')
if __name__ == '__main__':
run_test() |
|
Hi @SharkSyl, it'll work if you remove the |
ankane
added a commit
that referenced
this issue
Feb 9, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When running a SELECT or ADD query with cosine_distance on the embedding column, I encountered the following error:
The complete information is as follows：
My entire code is as follows：
The traceback:
The text was updated successfully, but these errors were encountered: