Deleting a dataset

If you have a dataset that is empty or a test dataset that you want to delete, you can perform a delete operation on the dataset. These two are the only scenarios where you can safely delete a dataset.

Datasets may contain some information representing data sources that are now part of the production or golden dataset. In such cases, deleting a dataset may seriously affect the health of your CMDB.

Before you begin

Before deleting a dataset, you must run a reconciliation job to purge or delete configuration items (CIs) from the dataset. Otherwise, the server returns an error when you try to delete the dataset.

For more information about all these normalization and reconciliation activities, see Normalizing data and Reconciling data.

To delete a dataset

1. To produce good identification results in reconciliation, use the Normalization Engine to normalize the CI attributes between the two datasets. 
   You perform this step to make it easier to determine which CIs represent the same object. If you do not intend to preserve the CIs, you can ignore this step.
2. For CIs that you want to preserve from the dataset that you intend to delete, create a standard Identification and Merge job to merge the CIs into the BMC.ASSET dataset.
   If you do not intend to preserve the CIs, you can ignore this step.
3. To remove all instances from the dataset, create and run a manual reconciliation job with a Delete or Purge activity to remove all the CIs from the dataset.
4. Open the BMC.CORE.CONFIG:BMC_Dataset form and search for the dataset record that you want to delete.
5. Delete the dataset record.
   The dataset is now removed from the server.