Delta Sharing import
Point, click, filter, import. Exponam Analyst Intelligence reads Delta Lake tables directly from storage through Delta Sharing and writes them to your worksheet. No SQL is written, no SQL warehouse runs, and no platform compute is metered. Available for Databricks.
Browse, filter, import
The Share View lists every table you are permitted to see, arranged by share and schema. Click a table to preview it. Filter any column – text, number, or date range – and check the columns you want. Save the configuration as a favorite and it appears under the table name for next time. Import to a new or current worksheet, with a PivotTable on the same sheet or a separate one if you want it, and refresh on demand or on a schedule.
Visual data builder
Share View, table preview, column selection, and filters.
What you need
Excel for Windows, Mac, or the web
This is the one path available on all three. The add-in installs from the public install page; name and email are requested; no payment.
A .share file
Issued by your Databricks administrator. It carries either an embedded credential (v1) or a pointer to your identity provider (v2), in which case you sign in through Microsoft Entra or Okta on first connection. A Databricks workspace account is not required, which is what makes this path usable for people outside the workspace.
Delta Sharing enabled in your Databricks environment
Standard in most enterprise deployments, and subject to internal approval. Where it has not been approved, direct SQL reaches the same tables through the SQL warehouse. Share and recipient setup is covered in the Getting started guide.
Why there is no compute
The request
The add-in asks the Delta Sharing server for a table, with your column selection and filters attached. Unity Catalog checks that the recipient in your .share file is entitled to that table and answers with a list of short-lived signed URLs pointing at the Parquet files behind it.
The read
The add-in reads those files straight from object storage and writes the rows to Excel. No SQL warehouse starts, no cluster is involved, and no DBUs are consumed. Because Parquet is compressed and columnar, only the columns and partitions you asked for cross the wire. On a standard Windows PC, one million rows land in about eleven seconds.
The warehouse never runs, so there is nothing to meter.
Cost and governance
What it costs
Nothing on the platform’s compute meter. The read comes from object storage under your cloud provider’s ordinary storage terms, the same as any other Delta Sharing recipient. Nothing routes through Exponam.
Governance
Access is governed by Unity Catalog. The share your administrator defines is the whole of what a recipient can see; a table outside the share cannot be reached from Excel. Each import can carry a properties sheet recording the server, the filter applied, the extraction time, and the license, so a workbook states where its numbers came from.
For the perimeter question, our Security Center documents data flow and the NIST SP 800-218 attestation.
Trying it
This is the path you can try before your administrator issues anything. The add-in ships with a sample Delta Share; install, open Import Data, and the sample tables are already in the Share View. Results land at up to one million rows per query without a license; enterprise licensing extends the ceiling to ten million rows and adds MSI deployment.
The other two paths: Natural-language query and Direct SQL.