Removing Dataset Outputs from Magic ETL

cookx775 · 2024-11-26T16:56:03+00:00

There was an error rendering this rich post.

We have been cleaning up our Domo instance removing unnecessary outputs from Magic ETL flows. I am assuming these datasets aren't truly being deleted, just disconnected from their source to not get updated any more. Is there a way to find these "orphaned" outputs so that they can be properly deleted?

Find more posts tagged with

Dataset

Magic ETL

Data Flows

Accepted answers

ArborRose

I think DomoStats has:

- Dataset ID
- Dataset Name
- Dataflow Name (if the dataset is used as part of a dataflow).
- Input/Output Type (whether the dataset is an input or output for a given dataflow).

All comments

ArborRose

Orphaned datasets would be datasets that, when investigated in DomoStats, would not have "In Use in Cards" or "In Use in Dataflows". So you should be able to use DomoStats Connector to connect to DomoStats -Dataset Usage and filter for no references in cards or dataflows, especially with a long "last accessed" timestamp.

CASE

  WHEN `In Use in Cards` = 0 AND `In Use in DataFlows` = 0 THEN 'Orphaned'

  ELSE 'In Use'

END

cookx775

When I am looking at "in use in dataflows" it seems to be pulling dataflows that are inputs for flows instead of both inputs and outputs.

What I am looking for are datasets that no longer have a dataflow associated with them at all (aka no dataflow is connected to power the dataset). I suppose I misused the term orphaned in this context as the number of cards connected is irrelevant.

Put differently, is there a way to see the input dataflow for my datasets in a report?

ArborRose

cookx775

Awesome, I was able to join the DataFlow Output Sources from domostats with Datasets from domostats to identify what I need (datasets from import_type = dataflow and Dataflow ID is null)