Overview

In complex data environments, data flows through multiple systems and technologies. In some cases, lineage might be incomplete, or contain deduced or duplicate assets. When you notice inaccuracies in your lineage, you can review aliases. Aliases serve as identifiers for specific systems and are displayed as connection strings, such as xyz-1ab2c3-x.ibm.com. When it is not possible to automatically map connections, aliases are identified and potential matching systems are suggested. The automatic process of searching for matching systems for a specific alias is called automapping. Some aliases might have only one match, while others might have multiple matches. To ensure accurate lineage, review these suggestions and assign each alias to the correct system (see Aliases to review ).

For example, you import data from Microsoft Power BI (Azure) and Microsoft SQL Server, and you know that two assets are connected. However, when you view lineage, the Microsoft Power BI (Azure) asset is linked to a deduced asset instead of the imported Microsoft SQL Server asset. In this situation, review the aliases that were automatically detected. You might find that an alias for the Microsoft Power BI (Azure) was identified, and a suggested matching system is a Microsoft SQL Server asset. Review the suggestion, and if it is correct, assign the alias to the matching system. Then, rerun the metadata import for the Microsoft Power BI (Azure) data source and review the lineage again. When the metadata import is complete, the Microsoft Power BI (Azure) asset is correctly connected to the Microsoft SQL Server asset.

You access aliases to review in the Data > Data lineage > Assign aliases tab.

After you assign aliases to suggested systems, you must rerun a metadata import for the changes to be visible on the lineage graph. To accelerate reimporting metadata, you can configure the metadata import to run only the analysis lineage phases. However, after reimporting, configure the metadata import job to include all lineage phases that were selected when the job was created.
In some cases, a suggested connection between an alias and one matching system is displayed on the list of aliases to review, but the lineage is already complete. In this case, when you assign the alias to the matching system, you don't need to reimport metadata and the lineage status is automatically set to Updated.

Aliases to review

Even if lineage might seem correct after the initial scan of automapping, adding more sources may cause some of the automapped aliases to stop working because of duplicated database/schema/table name structures in multiple database servers. Therefore, always confirm the suggested aliases to ensure identical results in the future. To review aliases, go to Data > Data lineage > Assign aliases and check Aliases to review section. An alias assignment candidate has the following elements:

• Alias name, which is a connection string, usually in the format similar to xyz-1ab2c3-x.ibm.com.
• Technology type of the alias.
• The number of occurrences, which is the number of inputs that contain the combination of alias connection string and alias type.
• The number of systems that use the alias, where a system is identified by the data source definition and the scanner type.
• System to assign, which is a candidate matching system. There might be one, multiple or no matching systems found.

When you review the aliases, start with those aliases for which the occurrences value is the highest. When you assign a system to this alias, it might resolve multiple lineage gaps.

Assigning aliases

To assign aliases and update the lineage, complete these steps:

1. Go to Data > Data lineage > Assign aliases.
2. On the Aliases to review tab, click an alias and review the suggested matching systems.
3. When you select the appropriate matching system, click Assign system.
4. Go to Reviewed aliases > Assigned section. The lineage status for the newly updated alias is Metadata import required. Display the details of a selected asset to see which systems require reimporting metadata.
5. When you identify which systems require metadata import, start reimporting metadata in the appropriate projects. For details, see Rerunning the import.
6. When reimporting finishes, view the updated lineage graph in Data > Data lineage.

Disabling alias automapping

In cases when you intentionally do not plan to scan the target/source and therefore not perform automapping to avoid potentially incorrectly mapped assets, as well as reduce the performance impact of automapping, you can disable it for specific aliases. For example, when some ETL jobs you scanned reference databases/sources that you are not interested in scanning for your current project, you can disable the automapping for them so that they are not reported for review. When you disable alias automapping, the alias is displayed on the Disabled aliases list, matching systems are not searched for it, and it is displayed as a deduced asset on the lineage graph. The disabled alias is not displayed on the list of aliases to review.

To disable automapping for an alias, complete these steps:

1. Go to Data > Data lineage > Assign aliases.
2. On the Aliases to review tab, click an alias.
3. Select the Disable alias automapping option and save the changes.

Learn more

• Viewing data lineage in Manta Data Lineage
• Importing metadata with Manta Data Lineage

Parent topic: Data lineage in Manta Data Lineage