Best Of

Here's the finalized forum response — ready to copy/paste:

Re-adding an orphaned output dataset via the Domo API

Option 3 from the accepted answer is the way to go if you want to preserve the historical data in the orphaned dataset. Here are the actual API requests:

Prerequisites: You'll need a Domo access token (Admin → Security → Access Tokens, or via the Developer Portal). You also need the dataset ID of the orphaned dataset (find it in the Data Center URL or via GET /api/v1/datasets).

Step 1 — GET the current dataflow definition

curl --request GET --url "https://{instance}.domo.com/api/dataprocessing/v1/dataflows/{dataflow_id}" --header "x-domo-authentication: {access_token}" --header "Content-Type: application/json"

This returns the full dataflow definition including an actions array (the ETL tiles) and an outputs array. Save this response — you'll need to modify it.

Step 2 — Add a new output tile in the UI

In the Domo UI, add a new Output DataSet tile to your dataflow. This creates a fresh dataset and a PublishToVault action in the definition. Don't worry about the name — you'll swap it out in the next step.

Save and close the dataflow editor.

Step 3 — GET the updated definition (with the new output)

curl --request GET --url "https://{instance}.domo.com/api/dataprocessing/v1/dataflows/{dataflow_id}" --header "x-domo-authentication: {access_token}" --header "Content-Type: application/json"

Find the new PublishToVault action in the actions array. It'll look something like:

{
"type": "PublishToVault",
"id": "abc123-...",
"name": "New Output DataSet",
"dependsOn": ["previous-tile-id"],
"dataSource": {
"guid": "NEW-DATASET-UUID",
"name": "New Output DataSet",
"type": "DataFlow"
},
"versionChainType": "REPLACE",
"schemaSource": "DATAFLOW",
"partitioned": false
}

Step 4 — Swap the dataset GUID to your orphaned dataset

Replace dataSource.guid with your orphaned dataset ID, and update the dataSource.name to match. Also update the outputs array at the top level of the definition — replace the new dataset ID with the orphaned one:

"outputs": [
{
"dataSourceId": "ORPHANED-DATASET-UUID",
"dataSourceName": "Your Original Output Name",
"versionChainType": "REPLACE"
}
]

Step 5 — PUT the updated definition

curl --request PUT --url "https://{instance}.domo.com/api/dataprocessing/v1/dataflows/{dataflow_id}" --header "x-domo-authentication: {access_token}" --header "Content-Type: application/json" --data @updated_dataflow_definition.json

Step 6 — Run the dataflow

curl --request POST --url "https://{instance}.domo.com/api/dataprocessing/v1/dataflows/{dataflow_id}/executions" --header "x-domo-authentication: {access_token}"

After the run completes, your orphaned dataset will be back in the dataflow as an output, and any cards built on that dataset will start refreshing again.

:warning: Warnings:
• Back up the original definition before modifying it (save the GET response from Step 1).
• The actions array must remain valid — don't break any dependsOn chains.
• Test on a sandbox dataflow first if you can.
• This approach works for Magic ETL dataflows. For MySQL/Redshift dataflows the structure differs slightly.

Want to discuss Domo questions like this in real time? Join the Domo User Group Slak

domousergroup.carrd.co

https://domousergroup.slack.com/archives/C07ESFPKAJ0/p1787757923655489

You have three options:

1 - Revert your ETL to a prior version that had these versions in it on the Versions tab of your dataflow. If you made additional changes to the ETL after removing these datasets, you'll have to reapply your changes.

2 - Delete the old historical dataset and just create new testing/validation datasets

3 - Add new output datasets to your ETL and then hack the JSON definition of the dataset to change the output to the old dataset IDs. This is a far more complex and risky route as you may end up invalidating your ETL and requires understanding API definitions and concepts.