retry function in dbt-core #7020
Conversation
|
Thank you for your pull request! We could not find a changelog entry for this change. For details on how to document a change, see the contributing guide. |
There was a problem hiding this comment.
very cool proof of concept :)
this command should probably not parse any project file, it should just read the manifest in that state
Yes, we need to do this to enable "manifest-only" execution! Related to many of the conversations we've been having. Where/how should we persist runtime config? In the manifest? As contents of raw dbt_project.yml file, or resolved Project object?
| ctx.obj["manifest"], | ||
| ) | ||
|
|
||
| from dbt import selected_resources |
There was a problem hiding this comment.
We need to de-globalize this (ik yk): #6009
| task.compile_manifest() | ||
| selected_resources.set_selected_resources(set(unique_ids)) | ||
| new_graph = task.graph.get_subset_graph(set(unique_ids)) | ||
| # should we give a way here for consumers to mutate the graph? |
There was a problem hiding this comment.
Maybe? Mutate the graph in the sense of:
- pass a
graphinto the task (I think this might be possible if we did [CT-1891] Abstract graph generation from task classes #6708) — a cached graph for now, but maybe someday a custom/modifiedgraph - define custom selectors / selection methods (a very cool & compelling idea!)
| if result.status not in [RunStatus.Success, TestStatus.Pass] | ||
| ] | ||
| # Using build task since it has a more complete runner set now | ||
| task = BuildTask( |
There was a problem hiding this comment.
Am I right in thinking - we'd need a way to dynamically load the appropriate task, based on previous_state.results.args.which?
|
Closing for now, in favor of the new issue: |
This code is garbage, just to demo some ideas RN
implement a retry command that can run your project from an existing state
TODO
Previous run
running retry
