Databend local mode Python Binding

Python binding for Databend in local mode - The multi-modal data warehouse built for the AI era.

Databend unifies structured data, JSON documents, and vector embeddings in a single platform with near 100% Snowflake compatibility. Built in Rust with MPP architecture and S3-native storage for cloud-scale analytics.

Installation

pip install databend

To test, run:

python3 -c "import databend; ctx = databend.SessionContext(); ctx.sql('SELECT version() AS version').show()"

API Reference

Core Operations

Method	Description
`connect(path=":memory:")`	Create a local embedded connection
`SessionContext()`	Create a new session context
`sql(query)`	Execute SQL query, returns DataFrame
`execute(query)`	Execute SQL and return the connection
`table(name)`	Query a registered table or view
`register(name, source)`	Register a local path, pandas/polars frame, or Arrow table
`from_df(obj, name=None)`	Materialize an in-process dataframe-like object as a relation

File Registration

Method	Description
`register_parquet(name, path, pattern=None, connection=None)`	Register Parquet files as table
`register_csv(name, path, pattern=None, connection=None)`	Register CSV files as table
`register_ndjson(name, path, pattern=None, connection=None)`	Register NDJSON files as table
`register_text(name, path, pattern=None, connection=None)`	Register TEXT files as table

Cloud Storage Connections

Method	Description
`create_s3_connection(name, key, secret, endpoint=None, region=None)`	Create S3 connection
`create_azblob_connection(name, url, account, key)`	Create Azure Blob connection
`create_gcs_connection(name, url, credential)`	Create Google Cloud connection
`list_connections()`	List all connections
`describe_connection(name)`	Show connection details
`drop_connection(name)`	Remove connection

Stage Management

Method	Description
`create_stage(name, url, connection)`	Create external stage
`show_stages()`	List all stages
`list_stages(stage_name)`	List files in stage
`describe_stage(name)`	Show stage details
`drop_stage(name)`	Remove stage

DataFrame Operations

Method	Description
`collect()`	Execute and collect results
`show(num=20)`	Display results in console
`to_pandas()`	Convert to pandas DataFrame
`to_polars()`	Convert to polars DataFrame
`to_arrow_table()`	Convert to PyArrow Table
`df()`	DuckDB-style alias for `to_pandas()`
`pl()`	Alias for `to_polars()`
`arrow()`	Alias for `to_arrow_table()`
`fetchall()`	Collect result rows as Python tuples
`fetchone()`	Return the first row or `None`

Examples

Local Tables

import databend
ctx = databend.connect()

# Create and query in-memory tables
ctx.execute("CREATE TABLE users (id INT, name STRING, age INT)")
ctx.execute("INSERT INTO users VALUES (1, 'Alice', 25), (2, 'Bob', 30)")
df = ctx.sql("SELECT * FROM users WHERE age > 25").df()

Working with Local Files

import databend
ctx = databend.connect("./demo-data")

# Query local Parquet files
ctx.read_parquet("/path/to/orders/", name="orders")
ctx.register("customers", "/path/to/customers.parquet")
df = ctx.sql("SELECT * FROM orders JOIN customers ON orders.customer_id = customers.id").df()

Working with In-Process DataFrames

import databend
import pandas as pd

ctx = databend.connect("./memory-store")
docs = pd.DataFrame(
    {
        "id": [1, 2],
        "content": ["hello", "vector memory"],
    }
)

ctx.register("docs", docs)
rows = ctx.sql("SELECT id, content FROM docs ORDER BY id").fetchall()

Cloud Storage - S3 Files

import databend
import os
ctx = databend.SessionContext()

# Connect to S3 and query remote files
ctx.create_s3_connection("s3", os.getenv("AWS_ACCESS_KEY_ID"), os.getenv("AWS_SECRET_ACCESS_KEY"))
ctx.register_parquet("trips", "s3://bucket/trips/", connection="s3")
df = ctx.sql("SELECT COUNT(*) FROM trips").to_pandas()

Cloud Storage - S3 Tables

import databend
import os
ctx = databend.SessionContext()

# Create S3 connection and persistent table
ctx.create_s3_connection("s3", os.getenv("AWS_ACCESS_KEY_ID"), os.getenv("AWS_SECRET_ACCESS_KEY"))
ctx.sql("CREATE TABLE s3_table (id INT, name STRING) 's3://bucket/table/' CONNECTION=(CONNECTION_NAME='s3')").collect()
df = ctx.sql("SELECT * FROM s3_table").to_pandas()

Development

# Setup environment
uv sync
source .venv/bin/activate

# Run tests
uvx maturin develop -E test
pytest tests/

Local Publish Without Docker

cd src/bendpy

# Build only
./scripts/local_publish.sh --python python3.12  --skip-upload

# Build only and clean old artifacts first
./scripts/local_publish.sh --python python3.12 --clean --skip-upload

# Build multiple Linux targets on the current host
./scripts/local_publish.sh --python python3.12 --clean --skip-upload \
  --target x86_64-unknown-linux-gnu \
  --target aarch64-unknown-linux-gnu

# Build and upload to PyPI
PYPI_TOKEN=your-token ./scripts/local_publish.sh --python python3.12

# Optional: override the package version for this build
PYPI_TOKEN=your-token ./scripts/local_publish.sh --python python3.12 --version 0.1.1

This script builds a wheel on the current host with maturin and uploads it with twine. It does not use Docker and does not produce a manylinux wheel. Cross-target builds also depend on the host having a working linker/toolchain for the requested target.

Metadata

Release files for databend 1.2.895

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Built distributions (wheels)

Table of built distributions (wheels) for databend 1.2.895
File	Interpreter	ABI	Platform	Reset
databend-1.2.895-cp312-abi3-manylinux_2_28_x86_64.whl 86.2 MB Apr 12, 2026 Without provenance	CPython 3.12	abi3	Linux glibc 2.28+ x86-64	Details
databend-1.2.895-cp312-abi3-manylinux_2_28_aarch64.whl 79.1 MB Apr 12, 2026 Without provenance	CPython 3.12	abi3	Linux glibc 2.28+ ARM64	Details
databend-1.2.895-cp312-abi3-macosx_11_0_arm64.whl 75.3 MB Apr 12, 2026 Without provenance	CPython 3.12	abi3	macOS 11.0+ ARM64	Details
databend-1.2.895-cp312-abi3-macosx_10_12_x86_64.whl 77.4 MB Apr 12, 2026 Without provenance	CPython 3.12	abi3	macOS 10.12+ x86-64	Details

Total release size: 318.1 MB

Release files / databend-1.2.895-cp312-abi3-manylinux_2_28_x86_64.whl

Download URL	databend-1.2.895-cp312-abi3-manylinux_2_28_x86_64.whl
Size	86.2 MB
Tags	CPython 3.12 Linux glibc 2.28+ x86-64 abi3
SHA-256 checksum How to use checksums	`d9de88cd999cacea39e66d83bb7215b7ae08937f95e89624fa708f691fcee469`
BLAKE2b-256 checksum How to use checksums	`63e7db155c6433cc74105c2c401e2e72286c01dd2ab4d2de8d716ea882ec1350`
Upload date	Apr 12, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.3`

Release files / databend-1.2.895-cp312-abi3-manylinux_2_28_aarch64.whl

Download URL	databend-1.2.895-cp312-abi3-manylinux_2_28_aarch64.whl
Size	79.1 MB
Tags	CPython 3.12 Linux glibc 2.28+ ARM64 abi3
SHA-256 checksum How to use checksums	`32951c570afe4878dc4c261d785b7122e25abb49d9159dbe35d272fb5d93f2b9`
BLAKE2b-256 checksum How to use checksums	`52aa62616f04ad114a089611c36d6d3f33d1b5806f1e919a208fe69ce6bc53ae`
Upload date	Apr 12, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.3`

Release files / databend-1.2.895-cp312-abi3-macosx_11_0_arm64.whl

Download URL	databend-1.2.895-cp312-abi3-macosx_11_0_arm64.whl
Size	75.3 MB
Tags	CPython 3.12 abi3 macOS 11.0+ ARM64
SHA-256 checksum How to use checksums	`d909b6e3b246b2e6c00ef8cf22a9e571605738482f6d8b5671462055df1a8ad5`
BLAKE2b-256 checksum How to use checksums	`8506007f78a7d0fcde2efa9fe576371a096da6465c5316c7325670190d4a6471`
Upload date	Apr 12, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.3`

Release files / databend-1.2.895-cp312-abi3-macosx_10_12_x86_64.whl

Download URL	databend-1.2.895-cp312-abi3-macosx_10_12_x86_64.whl
Size	77.4 MB
Tags	CPython 3.12 abi3 macOS 10.12+ x86-64
SHA-256 checksum How to use checksums	`5979bbb0725ac5396d0b14e963b2e1627e8e132b0ac5ff52565bc737f7a6b198`
BLAKE2b-256 checksum How to use checksums	`1a0c21dc32a9e71bda49c57f1761dd433f5bb4b3357b56ada9818c179dff958d`
Upload date	Apr 12, 2026
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/6.2.0 CPython/3.12.3`

Security for databend 1.2.895

Report project as malware

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.