Quickstart

This guide walks you through installing the SDK, authenticating, and running your first query.

Prerequisites

• Python 3.11, 3.12 or 3.13 (>=3.11, <3.14)

Step 1: Install

In a new directory, initialize a new uv project:

uv init --python "python >=3.11, <3.14"

Then, install the SDK:

uv add komodo

Verify installation:

uv run komodo --help

Step 2: Authenticate

Start the OAuth login flow:

uv run komodo login

Follow the browser prompt to authenticate with your Komodo credentials.

Step 3: Set Your Account

uv run komodo account set

Select your account from the interactive list.

Step 4: Run Your First Query

Create a file first_query.py:

import warnings
from komodo import get_snowflake_connection

# Suppress pandas warning about non-SQLAlchemy connections
warnings.filterwarnings("ignore", message="pandas only supports SQLAlchemy connectable")

# The SDK can also get a Snowflake connection and execute queries.
# This is a simple DBAPI2.0 wrapper that communicates with snowflake via proxy.
# Get the connection and cursor


def demo_get_all():
    print("Fetching all at once")
    conn = get_snowflake_connection()
    cursor = conn.cursor()
    cursor.execute("USE DATABASE DATA")
    query = "SELECT column_name, table_name FROM INFORMATION_SCHEMA.COLUMNS;"

    # Fetch the result set from the cursor and deliver it as the pandas DataFrame.
    cursor.execute(query)
    df_all = cursor.fetch_pandas_all()
    # https://docs.snowflake.com/en/developer-guide/python-connector/python-connector-api#fetch_pandas_all
    print(df_all.head(20))
    print("total rows:", len(df_all))


def demo_get_batches():
    print("Fetching in batches")
    conn = get_snowflake_connection()
    cursor = conn.cursor()
    cursor.execute("USE DATABASE DATA")
    query = "SELECT column_name, table_name FROM INFORMATION_SCHEMA.COLUMNS;"

    cursor.execute(query)
    # Fetch the result set from the cursor and deliver it as the pandas DataFrame.
    # https://docs.snowflake.com/en/developer-guide/python-connector/python-connector-api#fetch_pandas_batches
    for df in cursor.fetch_pandas_batches():
        print("New batch:")
        print("total rows in batch:", len(df))
        print(df)


if __name__ == "__main__":
    demo_get_all()
    demo_get_batches()

Run it:

uv run first_query.py

What You Just Did

1. **Created a new uv project
2. Installed the SDK from PyPI
3. Authenticated using OAuth 2.0 device flow
4. Selected your Komodo account
5. Executed a SQL query against Snowflake
6. Retrieved results using standard DB-API 2.0 patterns

Next Steps

• Execute more complex queries
• Use pandas for data analysis
• Handle long-running queries with async

Tip

Your credentials are saved in ~/.komodo/credentials. You won’t need to login again until your token expires.

Have Feedback?

We’d love to hear from you! Submit feedback about the SDK, documentation, or your experience using:

uv run komodo feedback