You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Add a new option for environment_suffix_target: catalog.
The desired behavior is that non-prod virtual environments would dynamically create new catalogs/databases. For example, sqlmesh plan dev results in:
mydb__dev.sushi.customers
The rationale is that this makes development with downstream consumers easier (e.g. reporting tools). Many tools will reference objects by schema.object. So, currently, if I want to test how a report or dataset behaves before promoting a model change to prod, I need to temporarily change the reference name to the schema or object inside the tool. If only the catalog/database name changed, then it is just a matter of temporarily updating the connection (which can usually be an environment variable) to use a different catalog.
Database engine limitations would mean only a subset of them could support this option. Primarily Snowflake, but also MotherDuck, Redshift, MSSQL, possibly Databricks.
The text was updated successfully, but these errors were encountered:
Add a new option for
environment_suffix_target: catalog
.The desired behavior is that non-prod virtual environments would dynamically create new catalogs/databases. For example,
sqlmesh plan dev
results in:The rationale is that this makes development with downstream consumers easier (e.g. reporting tools). Many tools will reference objects by
schema.object
. So, currently, if I want to test how a report or dataset behaves before promoting a model change to prod, I need to temporarily change the reference name to the schema or object inside the tool. If only the catalog/database name changed, then it is just a matter of temporarily updating the connection (which can usually be an environment variable) to use a different catalog.Database engine limitations would mean only a subset of them could support this option. Primarily Snowflake, but also MotherDuck, Redshift, MSSQL, possibly Databricks.
The text was updated successfully, but these errors were encountered: