The Data Discovery wizard generates a sensitive data model when you run a data discovery job. You can perform incremental updates to your sensitive data model while you are creating it in the wizard and after you save it to the Library.
- To launch the Data Discovery wizard, click the Home tab, and then click Data Discovery.
- (Optional) If you haven't granted the Data Discovery role on your target database, do the following:
- If your target database is a DB system, click
Download Privilege Script,
datasafe_privileges.sqlscript to your local computer, and then run the script on your target database.
- If your target database is an Autonomous Database, run the
DS_TARGET_UTILPL/SQL package on your Autonomous Database.
- If your target database is a DB system, click Download Privilege Script, download the
- Select a target database, and click Continue.If your target database is not listed, click Register and register your target database.
- On the Select Sensitive Data Model page, do the following:
- Select Create.
- Enter a name for your sensitive data model or leave the default name provided by Data Discovery.
- Choose whether to collect sample data from the target database and show the samples along with the discovery result.
- Select the compartment to which you want the sensitive data model to belong.
- Click Continue.
- On the Select Schemas for Sensitive Data Discovery page, select the schemas that you want Data Discovery to search, and click Continue. To select all the schemas at once, select the check box to the left of the Schema Name column title.
- (Optional) Move the Expand All slider to the right to view all categories and sensitive types. You can also expand individual check boxes.
- Select the categories of sensitive types and/or the individual sensitive types that you want to use to discover sensitive data.
- (Optional) Create a sensitive type:
- Click Add.
- Fill out the Create Sensitive Type dialog box.
- Click Save
- (Optional) Select the Use non-dictionary referential relationships for sensitive column discovery check box.
- When you are ready to start the data discovery job, click Continue.
When the job is completed successfully, a status of FINISHED is displayed.
- If you need to temporarily stop the data discovery job, click Suspend. Click Resume to continue.
- If you need to terminate the data discovery job, click Abort.
- Click Continue.
- If you enabled the "non-dictionary referential relationships" option, review the sensitive columns discovered with this option, deselect the columns that you do not want to include in your sensitive data model, and click Continue.
- On the Sensitive Data Discovery Result page, finalize the sensitive data model (SDM) by doing the following:
- Expand all of the results by moving the slider or by expanding certain nodes.
- (Optional) Select Schema View to sort the results by schema and table name.
- Review the sensitive columns and statistics (number of sensitive columns, estimated data counts, and sample data if you chose to collect sample data). If a sensitive column does not have a check box, it means that it has a referential relationship to a discovered sensitive column.
- Deselect any sensitive columns that you do not want to include in your SDM.
- (Optional) Click Add to add new sensitive columns. In the dialog box, select one or more columns from the schemas, select a sensitive type that describes the selected columns, and click Add to Result.
- (Optional) Click Back, modify the selection of sensitive types, and rerun the data discovery job. Review the generated SDM again. Repeat until you feel the SDM is accurate and complete.
- To view the Sensitive Data Discovery report for the SDM, click Report; otherwise, click Exit.
- (Optional) If you want to mask the sensitive data using the SDM, perform the following steps:
- Click Continue to mask the data.The Select Target for Data Masking page is displayed.
- Select the target database that you want to mask, and click Continue. The Masking Policy page is displayed.
- Create a masking policy.
- Click Continue to mask the data.