📓 Fusion Diaries: Query comments, VS Code usability updates, and Snowflake Iceberg table materializations

📓 Fusion Diaries: Query comments, VS Code usability updates, and Snowflake Iceberg table materializations

Author: Anders Swanson, Senior DX Advocate at dbt Labs

We’re nine months into building the dbt Fusion engine, and just three weeks away from Coalesce — register if you haven’t!

The dbt Fusion engine team is cooking, as usual. I dare you to read this week’s diary entry and tell us otherwise.

Velocity

  • 45 issues closed across the dbt-fusion and internal repos (is:issue state:closed closed:>=2025-07-01 closed:<2025-07-08)
  • 8 new betas released (our release process is getting smoother!)

New features

  • Query comments
  • VS Code extension usability updates
  • Snowflake Iceberg table materialization

(Abridged) Release notes

There are eight releases, so I won’t go over each one. Check out dbt-fusion’s CHANGELOG for more information.

Big rocks: What shipped this week

Query comments

The dbt Fusion engine now supports query comments. If you define a query-comment in your dbt_project.yml, it will be added to (almost) all queries that dbt executes.

The notable exception: comments aren’t added to some metadata queries that dbt fires off, such as those that start with SHOW and DESCRIBE.

VS Code extension

The new visual Getting Started walkthrough has landed!

Article content

We’ve heard from many folks that it hasn’t been clear that you need a full, successful compile of your project before you can preview a CTE or see red squiggles on incorrect SQL. Additionally, while a compile is running, the language server can’t reliably catch errors.

To make this more intuitive, we’ve made the “dbt Extension” section of the status bar turn yellow whenever a compile is happening.

Article content

[Legacy] Snowflake Iceberg table materialization

You can now materialize tables in Iceberg format.

This is an initial milestone that we’re considering legacy in favor of the cleaner Catalog integration for managed Iceberg tables.

See the corresponding Work in progress section below for more information.

“Small rocks”

Outside of the big rocks, we shipped some long‑awaited UX improvements:

  • dbt deps: --upgrade and --lock options
  • [VSCE]: shortcut to extension logs
  • [Databricks]: compute‑per‑model config

🚧 Work in progress

Snowflake Iceberg

Coming in the next few weeks:

  • Dynamic Iceberg tables
  • dbt Catalog integration specification (catalogs.yml)

BigQuery materialized views

They’re already available in an initial form, but there’s still some polish to be done. We expect to announce it as a “big rock” next week.

🐉 Here be dragons

Last week, we announced near parity with dbt Core for state:modified. We’ve since rolled that back after discovering some internal bugs, and we’re doing more testing. We’ll push the changes back out once we’re confident in the fixes.

🏁 Made it to the meme

A new habit I’ve developed: every day when I log in, I fetch the latest from main and trigger the language server to compile on our internal data project by typing CMD+S.

It’s become my security blanket — for good reason. Our project is very large, but in a few short minutes I gain confidence that I’m starting from a solid, verified commit. And when I want to preview a model or view compiled SQL, I can do it incredibly quickly thanks to incremental compilation.

Article content


To view or add a comment, sign in

More articles by dbt Labs

Others also viewed

Explore content categories