bug(python): Using a different vector_column_name for storing Vectors causes search() to fails about lancedb HOT 2 CLOSED

This is an issue with how you are ingesting data, the issue mentions - TypeError: Query column emb must be a vector. Got list<item: double>.

I ran your code snippet and the schema is a simple list and not an pyarrow list.

You are trying to create a table using a custom schema w/o providing it ('vector' is the default vector field for implicit schema detection). Our API reference mentions this :

Data is converted to Arrow before being written to disk. For maximum control over how data is saved, either provide the PyArrow schema to convert to or else provide a PyArrow Table directly.

The below code runs fine, pls use this, additionally you can also ingest data via LanceModel , everything is mentioned in our API reference for development, we will be updating the docs once the team has bandwidth.

import lancedb
import pyarrow as pa

custom_schema = pa.schema(
    [
        pa.field("emb", pa.list_(pa.float32(), 2)),
        pa.field("price", pa.float32()),
        pa.field("item", pa.string()),
    ]
)

uri = "./sample-lancedb"
db = lancedb.connect(uri)

table = db.create_table(
    "my_table",
    data=[
        {"emb": [3.1, 4.1], "item": "foo", "price": 10.0},
        {"emb": [5.9, 26.5], "item": "bar", "price": 20.0},
    ],
    schema=custom_schema,
)

result = table.search([100, 100]).limit(2).to_pandas()
print(result)

from lancedb.

closing this for now, its not a bug @changhiskhan / @wjones127

from lancedb.