Use this page to deploy, run, check, truncate and drop Warehouse tables day to day. To build a table first, see Create a dimension or fact table and Build, test and finalize a transformation.
Deploying creates the table in the warehouse from its current definition. Until then, the table shows Pending in the tree.
Deploying also builds the table's access layer and reapplies its Mind grants, so access is ready as soon as the table exists. Sync a table with the database covers the structure side of the deploy.
A Dictionary has no Deploy in the tree. Open it instead: its data mapping step shows a Deploy button until the table is deployed.
From the table:
From the transformation:
To see what a run writes, see How Dimension and Fact runs write data.
A Dimension with Store History keeps its versions current through a history run. It normally runs with the table's workflow, and a run can also be started on its own for one source system and a date. Dimension history explains what a run writes and the reasons a history step stops.
For it to run, the table must be a deployed Dimension with Store History on, and the source system must have a transformation with its schedule on. When it starts, DataMind shows History transformation started on JobId: followed by the job's ID.
The start is refused with the reason when something is missing or busy:
The Transformation Runs table on Settings lists each run with its ID, Name, Status, Note, Rows, Error, Duration, User, Start, Finish and Snap Date. It refreshes by itself while the tab stays open: every two seconds while a run is active, and every ten seconds otherwise. Click Refresh runs in the table toolbar to reload the list yourself.
The Status column carries the run's state: Running, Finished, Failed, Aborted, Error, Running Error or Skipped. A skipped step is not an error; Skipped runs and restarts explains why runs skip or are refused.
The statistics strip at the top of Settings shows Size, Rows and Partitions. Click Refresh to reload them. The transformation's own Execution tab keeps a Run Log; Run a transformation and read the Run Log explains it.
Truncate removes the data of one source system from a Dimension or Fact table. The table itself stays.
When the truncate finishes, DataMind shows Partitions truncated successfully. The full dialog is described with the rest of the tab in Dimension and fact settings.
Drop removes the stored table but keeps its definition, so you can deploy it again.
When the drop finishes, DataMind shows The table has been dropped from the database. A dropped table is no longer deployed, so it shows Pending again.
Dropping a table does not remove what depends on it. Those objects break until you fix them or redeploy the table. Check the Dependants tab first.
Open the table and go to Dependants. The tab lists downstream objects grouped by kind. If nothing uses the table, you see No dependants found.
Objects you cannot open are summarised as, for example, "2 Minds you cannot access". Ask their owners before you change the table. Dependencies and dependants describes the full tab, and Explore and trace lineage shows the wider picture.
Deleting removes a table or a transformation for good. Delete Warehouse objects covers the single and batch flows, the dependants check, refusals and retries.
| Action | Permission |
|---|---|
| Deploy, truncate, drop and delete a transformation | Add/Modify Warehouse Objects |
| Run and abort transformations | Execute/Abort Warehouse Objects |
| Run history for a source system | Execute/Abort Warehouse Objects |
| Delete tables | Remove Warehouse Objects |
| Read statistics and runs | View Warehouse Objects |
For the full picture, see Warehouse permissions and row-level security.
Truncate removes the data of one source system, and for a Fact, the dates you pick. The table stays. Drop removes the stored table but keeps its definition, so you can deploy it again.
Execute Now appears only after the transformation is finalized. Finish validation first, as described in Build, test and finalize a transformation.
Click the stop button on the running row, either in the table's Settings tab, which stops the run without a confirmation, or in the transformation's Execution tab, where you confirm Abort Transformation Run.
Yes. Pick the source system, then pick only the dates from last week under Partition Dates.
The versioned rows, not the current ones. Main clears the current data, History clears the versioned rows, and All clears both. Dimension history explains what the versions hold.
No. Truncation cannot be undone, and the confirmation says so before you click Truncate.
Yes, while the tab stays open: every two seconds while a run is active, and every ten seconds otherwise. Refresh runs reloads the list on demand.
Yes. Deleting a transformation needs Add/Modify Warehouse Objects. Deleting a table needs Remove Warehouse Objects.