Databricks Datashare

Access Dune’s blockchain data and your own Dune tables in Databricks with Delta Sharing

Databricks Datashare shares tables with your Unity Catalog metastore through Databricks-to-Databricks Delta Sharing. You query them in Databricks like any other catalog.

You can receive two kinds of tables:

• Dune’s data: Dune’s curated tables, set up by Dune for your team.
• Your own tables: tables in your Dune namespace with change data feed enabled, registered with sync_datashare, for example from dbt. See dbt Connector + Datashare.

There is no Databricks trial.

​

How It Works

1. Dune copies each table once, then applies its changes every five minutes by default.
2. Dune shares one view per table with your metastore, in one Delta Sharing share for your team.
3. You create a catalog from the share and query the views.

​

Get Started

Databricks Datashare is an enterprise feature configured per team. To enable it, email [e-mail] or contact the enterprise team with your sharing identifier.

​

Find Your Sharing Identifier

The sharing identifier names the Unity Catalog metastore that receives the share. Its format is <cloud>:<region>:<metastore uuid>, for example aws:us-west-2:19a84bee-54bc-43a2-87de-023d0ec16016.

Run this in a notebook or the SQL editor of the workspace you will use to query the data:

SELECT CURRENT_METASTORE();

You can also copy it from the Shared with me tab in Catalog Explorer. See Access using sharing identifier in the Databricks documentation.

Dune only needs the sharing identifier. Never send tokens, passwords, or other credentials.

​

Access the Data

When Dune confirms the share, it appears in the Shared with me tab in Catalog Explorer. Create a catalog from it:

CREATE CATALOG IF NOT EXISTS dune USING SHARE <provider-name>.<share-name>;

This requires a metastore admin, or the CREATE CATALOG and USE PROVIDER privileges. See Read data shared using Databricks-to-Databricks Delta Sharing.

Each table appears as a view with the same schema and table name as in Dune:

SELECT *
FROM dune.ethereum.transactions
LIMIT 10;

​

Requirements

• A Databricks workspace enabled for Unity Catalog.
• A SQL warehouse or a cluster running Databricks Runtime 14.1 or above. The tables behind the shared views use deletion vectors, which older runtimes cannot read.
• Databricks limits on shared views apply: for example, at most 20 shared views in one query, and no history or streaming reads. See Read data shared using Databricks-to-Databricks Delta Sharing.

​

Data Types

varbinary columns (addresses, hashes) arrive as BINARY, not as hex strings. Filter on them with a binary literal:

WHERE tx_hash = X'efb2e2c26974f72d9f3f04c693db73ecc679dd60'

Databricks has no UINT256 or INT256 type. In Dune’s tables, Dune converts these columns to a supported type, for example DOUBLE. In your own tables, cast them before syncing, see Configuration.

​

Details

• Freshness: changes are applied every five minutes by default
• Pricing: contact the enterprise team
• Support: [e-mail] | [e-mail]