What Atlan crawls from Informatica PowerCenter
Atlan reads your Informatica PowerCenter repository XML exports and reconstructs the data flow through each mapping, session, and workflow. In its current release, the connector produces lineage only. It doesn't create PowerCenter catalog assets such as folders, mappings, or workflows in Atlan.
Lineage
Atlan statically analyzes the repository exports to trace how each target is populated, then publishes lineage at both table level and column level between your already-cataloged source and target assets.
Table-level lineage
Table-level lineage tracks data flow from source tables and files, through PowerCenter transformations, to target tables:
Oracle.SALES.CUSTOMERS → [PowerCenter mapping: m_load_dim_customer] → Snowflake.ANALYTICS.DIM_CUSTOMER
Column-level lineage
Column-level lineage tracks how individual target columns are derived from source columns, including pass-through columns and transformation expressions:
Oracle.SALES.CUSTOMERS.FIRST_NAME → [expression: concat] → Snowflake.ANALYTICS.DIM_CUSTOMER.FULL_NAME
Oracle.SALES.CUSTOMERS.LAST_NAME → [expression: concat] → Snowflake.ANALYTICS.DIM_CUSTOMER.FULL_NAME
Lineage assets
Atlan maps the PowerCenter data flow to its native lineage asset types.
| PowerCenter source | Atlan asset type | Description |
|---|---|---|
| Mapping and session data flow | Process | Table-level lineage between a set of source and target assets |
| Field-to-field mappings within a data flow | ColumnProcess | Column-level lineage between individual source and target columns |
Lineage assets connect to the source and target tables in Atlan. When a referenced table hasn't been cataloged yet, Atlan still publishes the lineage by creating a partial asset in its place.
Partial assets
The connector has explicit support for partial assets. When a PowerCenter mapping references a source or target that hasn't been crawled, mined, or cataloged in Atlan yet, Atlan still publishes the lineage by creating a partial asset as a placeholder.
If a supported data source has not been crawled, mined, or cataloged via APIs but is present in lineage, Atlan will create partial assets to provide a complete representation of data lineage.
This lets your PowerCenter lineage appear end-to-end even before every source and target system is fully cataloged. Key characteristics of partial assets include:
- Displayed with a dotted border on the lineage graph.
- You can search for and view a partial asset profile and lineage.
- You can view metadata in the sidebar, but you can't enrich a partial asset.
- Column-level lineage for partial assets is supported.
A partial asset resolves into a full, discoverable asset once you crawl the underlying source or target system. For more detail, see What are partial assets.
Supported source and target types
Atlan resolves the underlying database for each source and target from the PowerCenter connection metadata. The following relational types are recognized automatically:
Oracle, Microsoft SQL Server, IBM DB2, Sybase, Teradata, Netezza, Greenplum, PostgreSQL, Snowflake, and Amazon Redshift.
| Source type | Handling |
|---|---|
| Recognized relational databases (listed in this article) | Resolved automatically |
| Flat file, XML, and VSAM sources | Appear as file-type source or target endpoints in lineage |
| ODBC and other non-relational connections | May not resolve to a specific database in the current release; contact your Atlan representative if these are missing from lineage |
How Atlan resolves sources
To build accurate lineage, Atlan interprets the following PowerCenter constructs:
- Source qualifiers: the source table is resolved from its owner or schema metadata.
- SQL overrides: when a source qualifier defines a SQL override, that SQL is used as the source of record for lineage.
- Lookup transformations: the looked-up table and schema are resolved, including overridden lookup targets.
- Parameters: parameter tokens such as
$$Paramand$Paramare resolved from the parameter files you upload. Unresolved tokens are flagged as low-confidence.
See also
- Crawl Informatica PowerCenter assets: Configure and run the workflow
- Lineage: How lineage is calculated and known limitations