Status quo
Currently, table names must be in the format <schema>.<table>. This means dlt_run_params will assign those values to dataset_name, table_name.
Pitch
For generalization purposes, let's start addressing database resources by catalog.schema.table going forward.
Details
Assign those component values to dlt_run_params' project_name, dataset_name, table_name. For some databases like DuckDB, "catalog = attached database". For a full mapping of triples across different database dialects, see ingestr's new capabilities.go. The alternative project.dataset.table jargon is apparently used by BigQuery. While being at it, bundle the triple into a dataclass object and add corresponding parsing and validation methods instead of spreading them across the codebase.
References
Status quo
Currently, table names must be in the format
<schema>.<table>. This meansdlt_run_paramswill assign those values todataset_name,table_name.Pitch
For generalization purposes, let's start addressing database resources by
catalog.schema.tablegoing forward.Details
Assign those component values to
dlt_run_params'project_name,dataset_name,table_name. For some databases like DuckDB, "catalog = attached database". For a full mapping of triples across different database dialects, see ingestr's new capabilities.go. The alternativeproject.dataset.tablejargon is apparently used by BigQuery. While being at it, bundle the triple into a dataclass object and add corresponding parsing and validation methods instead of spreading them across the codebase.References