# Databricks

Sumble connects directly to your Databricks workspace to read your account data and write enriched results back as Delta tables.

## How it works

Sumble uses a direct SQL connection to your Databricks workspace (via the Databricks SQL API). The integration is bi-directional:

**Pull** — Sumble queries your account list from a table you specify (e.g., `main.customer_schema.accounts`). Column mappings translate your field names to Sumble's schema. This tells Sumble which organizations to match and enrich.

**Push** — After enrichment, Sumble writes the results as Delta tables in your Databricks workspace:

1. Parquet files are generated from the enrichment pipeline
2. Files are uploaded to intermediate cloud storage
3. Delta tables are created in a timestamped schema in your workspace
4. Tables are atomically swapped into your output schema, so you always see a complete, consistent dataset

## What gets delivered

Each pipeline run writes these tables to your output schema:

* Enriched organizations — your accounts matched to Sumble organizations with all configured enrichment columns
* Alerts feed — intent signals generated for your accounts
* Reference data — technologies, job functions, countries, and other lookup tables

Tables are replaced atomically with each run. Previous versions are retained briefly for rollback if needed.

## Update cadence

Data is refreshed daily after the Sumble data pipeline completes.

## Setup

Databricks integration requires an enterprise plan. To get started:

1. Provide Sumble with your Databricks workspace hostname, SQL warehouse HTTP path, and a Personal Access Token
2. Specify the input table containing your account list and the output schema for enriched data
3. Sumble configures column mappings for your account fields (account ID, name, URL, etc.)
4. Choose which enrichments to include

<a href="https://calendly.com/d/cnzk-sjk-q38/sumble" class="button primary">Book a time to chat with us</a>