LanceDB 中文文档

外观

本页尚未翻译

本页内容仍为英文原文，源自官方文档 integrations/data/pandas_and_pyarrow。中文翻译正在分批进行中。

Pandas and PyArrow ​

Applies to: Embedded ✅, Enterprise ✅

Because Lance is built on top of Apache Arrow, LanceDB fits naturally into Pandas-first workflows. You can ingest a DataFrame, query it with LanceDB's vector operators, and keep working in Pandas without any glue code.

Create a dataset ​

Start by importing LanceDB alongside your usual Pandas utilities and connect to a temporary database.

python
import asyncio
import tempfile
from pathlib import Path

import lancedb
import pandas as pd

Use the familiar pd.DataFrame API to prepare your rows, then pass the entire frame to db.create_table.

python
import lancedb

pandas_df = pd.DataFrame(
    [
        {"id": "1", "text": "dragon", "vector": [0.9, 0.1, 0.3]},
        {"id": "2", "text": "griffin", "vector": [0.4, 0.5, 0.2]},
        {"id": "3", "text": "phoenix", "vector": [0.7, 0.3, 0.6]},
    ]
)
pandas_db = lancedb.connect(str(Path(tempfile.mkdtemp()) / "pandas-demo"))
pandas_table = pandas_db.create_table("creatures", data=pandas_df, mode="overwrite")

Queries can return Pandas frames as well, so you can immediately inspect the results or pipe them into downstream analytics.

python
pandas_results = (
    pandas_table.search([0.9, 0.1, 0.3])
    .select(["text", "_distance"])
    .limit(1)
    .to_pandas()
)
print(pandas_results)

Async API ​

For web services or background jobs that already rely on asyncio, use the asynchronous helpers to keep everything non-blocking.

python
import lancedb

async def run_pandas_async_example() -> pd.DataFrame:
    async_db = await lancedb.connect_async(
        str(Path(tempfile.mkdtemp()) / "pandas-async")
    )
    async_df = pd.DataFrame(
        [
            {"id": "10", "text": "sage", "vector": [0.6, 0.4, 0.8]},
            {"id": "11", "text": "bard", "vector": [0.2, 0.7, 0.3]},
        ]
    )
    async_table = await async_db.create_table(
        "creatures_async", data=async_df, mode="overwrite"
    )
    # An async search returns its query builder from an await
    query = await async_table.search([0.6, 0.4, 0.8])
    async_results = await query.select(["text", "_distance"]).limit(1).to_pandas()
    print(async_results)
    return async_results

async_results = asyncio.run(run_pandas_async_example())