Skip to content

Overview of Tables in LanceDB

A Table is a collection of Records in a LanceDB Database. Tables in Lance have a schema that defines the columns and their types. These schemas can include nested columns and can evolve over time.

Key Concepts

  • Tables: The fundamental unit of data organization in LanceDB, containing records with a defined schema
  • Schema: Defines the structure of your data, including column names, data types, and whether fields are required
  • Vector Column: A special column type that stores vector embeddings for similarity search
  • Metadata Columns: Regular columns that store additional information about each record
  • Versioning: Tables maintain versions as they are modified, enabling point-in-time queries

Operations Overview

LanceDB provides comprehensive table management capabilities:

  • Creation: Create tables from various data sources including lists, DataFrames, and Pydantic models
  • Modification: Add, update, or delete data with support for batch operations
  • Schema Management: Add, alter, or drop columns as your data needs evolve
  • Consistency Control: Configure read consistency settings for your use case

Legacy Javascript SDK

The vectordb package is a legacy package that is deprecated in favor of @lancedb/lancedb. The vectordb package will continue to receive bug fixes and security updates until September 2024. We recommend all new projects use @lancedb/lancedb.