The companion piece for the people actually writing the SQL.
(more…)What every manager sponsoring AI analytics needs to understand before the project starts
When you remove a model from your dbt project, the table stays in your warehouse. We developed a macro that compares the current dbt project resources (models, seeds, and snapshots) with the objects that already exist in your database. The macro looks for existing database objects in the information schema, and it scans the current dbt project resources using the dbt graph. It then compares these two sets.
If dbt does not find a corresponding database object for a given dbt resource, it can generate or even execute a DROP command, allowing you to keep your database clean and free of unused objects.
In this article, we will explore a practical example of how this can be used in practice.
(more…)dbt State-Aware Orchestration (SAO) with the Fusion engine runs only what actually changed in your DAG — skipping unchanged models to save time and money. It detects code and data changes automatically, supports smart controls like build_after and updates_on, and delivers faster, more stable pipelines with up to ~29% cost savings.
Have you ever wondered how to use large language models on your own data? Or what becomes possible when you combine standard analytics tools with LLMs?
Applying LLMs in real-world analytics has historically been difficult. Not because the technology didn’t exist, but because cost, complexity, reliability, and governance created significant barriers to adoption.
This article explores why that integration has been challenging — and how modern data platforms, particularly Snowflake, are changing what becomes feasible when applying LLMs to practical analytics use cases.
