Lance REST integration

Overview

This guide provides comprehensive instructions for integrating the Apache Gravitino Lance REST service with data processing engines that support the Lance format, including Apache Spark via the Lance Spark connector and Ray via the Lance Ray connector.

This documentation assumes familiarity with the Lance REST service setup as described in the Lance REST Service documentation.

Compatibility matrix

Datastrato Enterprise 1.3.0 is verified with the connector versions below. Test the exact versions you plan to run in your own environment before production use, because Lance connector releases change quickly.

Connector	Supported with 1.3.0	Not supported
lance-spark	0.2.0, 0.4.0, 0.5.1	0.1.0 and 0.1.1: every table create fails with 404 Not Found
lance-ray	0.3.0, 0.4.2; 0.2.0 with pylance pinned to 3.x or 4.x	0.1.0: write_lance fails with TypeError on the namespace_impl argument

Reproducing the matrix locally

Both connectors ship with a multi-version integration test driver so the matrix can be re-verified (and extended) without ad-hoc scripting:

Run the connector version matrix tests

# lance-spark: runs LanceSparkRESTServiceIT once per bundle version.
# The default list intentionally omits 0.1.0 / 0.1.1: those bundles call the
# removed /create-empty endpoint and will fail with 404 against 1.3.0+.
./gradlew :lance:lance-rest-server:lanceSparkMatrixTest \
    -PlanceSparkBundleVersions=0.2.0,0.4.0,0.5.1 \
    -PskipDockerTests=true
# Per-version JUnit reports land under
# lance/lance-rest-server/build/reports/lance-spark-matrix/{version}/.

# lance-ray: provisions a venv per version under
# clients/client-python/build/lance-ray-matrix/.venv-{version}/ and runs
# tests/integration/test_lance_ray.py against each. The Gradle wrapper
# below starts / stops Gravitino automatically.
./gradlew :clients:client-python:lanceRayMatrixTest \
    -PlanceRayVersions=0.4.2,0.3.0

Rationale

The Lance ecosystem is under active development, with frequent updates to APIs and features. Gravitino's Lance REST service depends on specific connector behaviors to ensure reliable operation. Using incompatible versions may result in:

• Runtime errors or exceptions
• Data corruption or loss
• Unexpected behavior in query execution
• Performance degradation

Prerequisites

Before proceeding, ensure the following requirements are met:

1. Gravitino Server: A running Gravitino server instance with the Lance REST service enabled

   • Default endpoint: http://localhost:9101/lance

2. Lance Catalog: A Lance catalog created in Gravitino using either:

   • Lance REST namespace API (CreateNamespace operation - see Lance REST Service documentation
   • Gravitino REST API, for more, refer to lakehouse-generic-catalog
   • Example catalog name: lance_catalog

3. Lance Spark Bundle (for Spark integration):

   • Downloaded lance-spark bundle JAR matching your Apache Spark version
   • Note the absolute file path for configuration

4. Python Dependencies:

   • For Spark integration: pyspark
   • For Ray integration: ray, lance-namespace, lance-ray

Spark integration

Configuration

Configure PySpark and query Lance REST

from pyspark.sql import SparkSession
import os
import logging

# Configure logging for debugging
logging.basicConfig(level=logging.INFO)

# Configure Spark to use the lance-spark bundle
# Replace /path/to/lance-spark-bundle-3.5_2.12-X.X.XX.jar with your actual JAR path and version;
# refer to the compatibility matrix for supported lance-spark versions.
os.environ["PYSPARK_SUBMIT_ARGS"] = (
    "--jars /path/to/lance-spark-bundle-3.5_2.12-0.4.0.jar "
    "--conf \"spark.driver.extraJavaOptions=--add-opens=java.base/sun.nio.ch=ALL-UNNAMED\" "
    "--conf \"spark.executor.extraJavaOptions=--add-opens=java.base/sun.nio.ch=ALL-UNNAMED\" "
    "--master local[1] pyspark-shell"
)

# Initialize Spark session with Lance REST catalog configuration
# Note: The catalog "lance_catalog" must exist in Gravitino before running this code, you can create
# it via Lance REST API `CreateNamespace` or Gravitino REST API `CreateCatalog`.
spark = SparkSession.builder \
    .appName("lance_rest_integration") \
    .config("spark.sql.catalog.lance", "com.lancedb.lance.spark.LanceNamespaceSparkCatalog") \
    .config("spark.sql.catalog.lance.impl", "rest") \
    .config("spark.sql.catalog.lance.uri", "http://localhost:9101/lance") \
    .config("spark.sql.catalog.lance.parent", "lance_catalog") \
    .config("spark.sql.defaultCatalog", "lance") \
    .getOrCreate()

# Enable debug logging for troubleshooting
spark.sparkContext.setLogLevel("DEBUG")

# Create schema (database)
spark.sql("CREATE DATABASE IF NOT EXISTS sales")

# Create Lance table with explicit location
spark.sql("""
    CREATE TABLE sales.orders (
        id INT,
        score FLOAT
    )
    USING lance
    LOCATION '/tmp/sales/orders.lance/'
    TBLPROPERTIES ('format' = 'lance')
""")

# Insert sample data
spark.sql("INSERT INTO sales.orders VALUES (1, 1.1)")

# Query data
spark.sql("SELECT * FROM sales.orders").show()

Storage location configuration

The LOCATION clause in the CREATE TABLE statement is optional. When omitted, lance-spark automatically determines an appropriate storage location based on catalog properties. For detailed information on location resolution logic, refer to the Lakehouse Generic Catalog documentation.

For Gravitino-managed Lance catalogs, put the storage configuration in the Gravitino catalog properties so Spark does not need to repeat it.

Create a catalog with Lance storage properties

curl -X POST -H "Accept: application/vnd.gravitino.v1+json" \
  -H "Content-Type: application/json" \
  -d '{
  "name": "lance_catalog",
  "type": "RELATIONAL",
  "provider": "lakehouse-generic",
  "comment": "catalog for Lance tables on MinIO",
  "properties": {
    "location": "s3://bucket/tmp",
    "lance.storage.endpoint": "http://minio:9000",
    "lance.storage.access_key_id": "ak",
    "lance.storage.secret_access_key": "sk",
    "lance.storage.allow_http": "true",
    "lance.storage.region": "us-east-1"
  }
}' http://localhost:8090/api/metalakes/test/catalogs

Create a table under the catalog location

spark.sql("""
    CREATE TABLE sales.orders (
        id INT,
        score FLOAT
    )
    USING lance
    LOCATION 's3://bucket/tmp/sales/orders.lance/'
    TBLPROPERTIES ('format' = 'lance')
""")

If you need a per-table override, lance.storage.* table properties are still supported and take precedence over catalog defaults.

Ray integration

Installation

Install lance-ray

pip install lance-ray

• Ray will be automatically installed if not already present
• For Gravitino 1.3.0, use a lance-namespace client compatible with server-side lance-namespace-core 0.7.5 or newer.
• Ensure Ray version compatibility in your environment before deployment

Example

Read and write Lance tables with Ray

import ray
import lance_namespace as ln
from lance_ray import read_lance, write_lance

# Initialize Ray runtime
ray.init()

# Connect to Lance REST namespace
namespace = ln.connect("rest", {"uri":  "http://localhost:9101/lance"})

# Create sample dataset
data = ray.data.range(1000).map(
    lambda row: {"id": row["id"], "value": row["id"] * 2}
)

# Write dataset to Lance table
# Note: Both the catalog "lance_catalog" and schema "sales" must exist in Gravitino, you can create
# them via Lance REST API `CreateNamespace` or Gravitino REST API `CreateCatalog` and `CreateSchema`.
write_lance(
    data, 
    namespace=namespace, 
    table_id=["lance_catalog", "sales", "orders"]
)

# Read dataset from Lance table
ray_dataset = read_lance(
    namespace=namespace, 
    table_id=["lance_catalog", "sales", "orders"]
)

# Perform filtering operation
result = ray_dataset.filter(lambda row: row["value"] < 100).count()
print(f"Filtered row count: {result}")

Additional engines

The Lance REST service is compatible with other data processing engines that support the Lance format, including:

• DuckDB: For analytical SQL queries
• Pandas: For Python-based data manipulation
• DataFusion: For Rust-based query execution

Note: These three engines do not support Lance REST natively yet, but can still interact with Lance datasets through table location paths retrieved from the Lance REST service.

For engine-specific integration instructions, consult the Lance Integration Documentation.

General integration pattern

Most Lance-compatible engines follow this general pattern:

1. Establish connection to Lance REST service endpoint
2. Authenticate using appropriate credentials
3. Reference tables using the hierarchical namespace structure
4. Execute read/write operations using engine-native APIs

Refer to each engine's specific documentation for detailed configuration parameters and code examples.

Related

• Lance REST Service Documentation
• Lance Format Specification
• Apache Gravitino Documentation
• Lakehouse Generic Catalog Guide