rjh336 / dbt-model-usage Goto Github PK

What

As someone that looks at the usage of bigquery tables, it would be great to see the usage per column on a table.

Why

Since BigQuery is a columnar database, most users will select the data they need per column. Having information about usage a per column basis would help identify which columns in a table are no longer used.

How

I want to built a solution like this anyways so it would be great if we could collaborate on this and I could add it to this package.

These are two types of usage patterns of columns that I thought of. First, there is the non-specific 'SELECT *' that would mean the column gets queried but does not indicate that the users really wants a specific column. Second, there is the specific 'SELECT column' that indicates that the column is being used. This last one would also hold for cases where it is used in the other clauses like WHERE, ORDER BY, GROUP BY, and HAVING

Steps to reproduce:

Would be great to have redshift support for this

Expected resuts:

also works for redshift

Actual results

only bigquery and snowflake