Skip to main content
Community Hub

Crawl Informatica PowerCenter assets

TL;DR

Create a workflow to extract table-level and column-level lineage from Informatica PowerCenter repository exports.

Your AI can read this via Docs MCPcurl -fsSL "https://docs.atlan.com/install-docs-mcp" | bashConnect

Create a workflow in Atlan to extract lineage from Informatica PowerCenter repository exports. Each workflow processes the repository exports you upload or import from cloud storage and publishes the resulting lineage against a single Atlan connection. Review the order of operations for metadata enrichment workflows before starting.

Prerequisites​

Before you begin, make sure you have:

  • Completed Informatica PowerCenter setup
  • A .zip archive of your repository XML exports, ready to upload or staged in a cloud object store if you plan to import from object storage
  • A .zip archive of the parameter files your workflows run with. If your sessions use connection variables such as $DBConnection_SALES, include the files that assign them: without those values, Atlan can't tell which database the tables come from
  • (Recommended) the source and target systems already crawled in Atlan, so lineage resolves to real assets. If they aren't cataloged yet, Atlan still publishes lineage using partial assets that resolve automatically once you crawl those systems

Create crawler workflow​

  1. In your Atlan workspace, click Connectors in the left sidebar.
    • If you are using the Old UI (Classic), click New Workflow in the top navigation.
  2. Click Marketplace.
  3. Search for Informatica PowerCenter and select it.
  4. Click Install.
  5. Once installation completes, click Setup Workflow on the same tile.

Upload repository files​

  1. In the Repository XML export (.zip) field, upload the .zip archive containing your PowerCenter repository XML exports. This field is required.
  2. In the Parameter files (.zip) field, upload the .zip archive of parameter files. The field is optional, but most PowerCenter projects need it for accurate lineage. Include every file that assigns a connection variable, such as $DBConnection_SALES: Atlan uses the connection each variable resolves to.

Import from object storage​

Instead of uploading files, you can have the workflow fetch your repository exports directly from a cloud object store such as Amazon S3, Google Cloud Storage (GCS), or Azure Data Lake Storage (ADLS). This is useful for larger exports, automated pipelines, or recurring runs where the files are produced outside Atlan.

  1. Set Import type to Cloud storage.
  2. In Cloud source, select the object-store credential that holds your repository export. Leave it empty to read from your tenant's own object store. For the per-provider fields for S3, GCS, and ADLS, see Object storage configuration for apps.
  3. In Repository XML prefix, enter the object-store directory that holds the repository XML export. Alternatively, or in addition, set Repository XML key to the exact object key of the export .zip. At least one of the two is required.
  4. (Optional) In Parameter files prefix, enter the object-store prefix that holds your parameter files.

Under a prefix the workflow accepts either a .zip archive or loose .xml files. Keep filenames unique within a prefix: two files that resolve to the same name are rejected rather than silently overwritten.

Configure advanced options​

  1. Use the Operator overrides (JSON) field to give Atlan per-connection settings that the PowerCenter XMLs don't carry. The most common are naming the engine behind an ODBC connection, and mapping a PowerCenter connection to one Atlan connection when the same tables are cataloged in more than one. If lineage looks right, leave the field as {}.

The field holds a JSON object with a connections map. Each key is a PowerCenter connection name, as shown in Workflow Manager, and each value lists the settings for that connection:

{
"connections": {
"ODBC_PROD": {"dbtype": "Snowflake"},
"SALES_DB2": {
"matchConnection": [{"attribute": "name", "value": "sales-prod"}]
}
}
}

Connection names match regardless of case. Give each connection only one entry: two entries that differ only in case are rejected.

Use the real connection name, not a connection variable such as $DBConnection_SALES. A variable can point to a different connection in each workflow, so a setting keyed on it can apply to the wrong tables. Upload the parameter file that assigns the variable, and use the connection name it resolves to. See Partial assets from unresolved connection variables.

Connection settings​

SettingTypeWhat it does
dbtypestringThe engine the connection really talks to, as a PowerCenter database type: for example Snowflake, Oracle, Microsoft SQL Server, Redshift, PostgreSQL, Teradata, or DB2. Required for ODBC connections, because ODBC names a protocol, not a database. Replaces the engine the XMLs declare.
matchConnectionlist of rulesTies the PowerCenter connection to an Atlan connection. See Map connection to Atlan connection.
databaseNamestringThe database name to use for the connection's tables. Always replaces a database name from the XMLs.
schemaNamestringThe schema to use when the XMLs don't name one, for example a target with a blank Table Name Prefix or SQL that names a bare table. A schema in the XMLs always wins.
dbdnameIsCatalogtrue or falseWhether to use the database name stored in the PowerCenter source or target definition. By default Atlan uses it only for engines where it reliably names the physical database (Microsoft SQL Server, PostgreSQL, Netezza, Greenplum, Snowflake, and Redshift). Set false if your definitions store an alias instead, or true to use it for another engine.
notesstringA free-text note for whoever maintains the overrides. Atlan ignores it.

Map connection to Atlan connection​

PowerCenter only knows its own connection names, so when the same tables are cataloged in more than one Atlan connection, such as a production database and a replica, Atlan can't tell which copy a mapping reads or writes. Rather than guess, it creates partial assets. A matchConnection mapping makes the choice explicit:

{"connections": {"SALES_DB2": {"matchConnection": [{"attribute": "name", "value": "sales-prod"}]}}}

Each rule matches one attribute of the Atlan connection:

attributeMatchesUse it when
nameThe Atlan connection's name, ignoring caseYou want exactly one Atlan connection. The name must be unique on your tenant.
hostThe host the Atlan connection was crawled fromEach database server is cataloged as one Atlan connection

If you give several rules, an Atlan connection must match all of them.

warning

A mapping is strict. Every table of the mapped PowerCenter connection links only to tables in the matched Atlan connection, even if the same table name exists in other connections. A table that isn't found there appears as a partial asset. It's never linked to a table in another connection.

Apply settings to every connection​

A "*" entry applies to every PowerCenter connection that doesn't have an entry of its own:

{
"connections": {
"*": {"matchConnection": [{"attribute": "name", "value": "warehouse-prod"}]},
"MART_CONN": {
"schemaName": "mart",
"matchConnection": [{"attribute": "name", "value": "warehouse-prod"}]
}
}
}
  • The "*" entry covers connections on every engine. A connection that talks to a different engine than the "*" target needs its own entry, or its tables won't be found.
  • A connection's own entry replaces "*" completely: settings aren't combined. In the example, MART_CONN repeats the matchConnection rule, because without it the connection has no mapping.
  • The "*" entry can't set an engine: dbtype isn't allowed there. Set the engine on each connection's own entry.

Workflow-wide settings​

SettingTypeWhat it does
emitPartialIdentitytrue or falseDefault true. When a table is named without a database or schema, Atlan matches it on the parts it has. Set false to turn this off for the whole workflow. This also stops Atlan from using any database name from the PowerCenter definitions, and overrides dbdnameIsCatalog: true.

Configure connection​

  1. Select an existing Connection or create a new one. This connection becomes the root for all lineage assets Atlan publishes from this workflow.
  2. To change who manages this connection, update the users or groups listed under Connection admins.
info

This connector doesn't run a test authentication or preflight connectivity check. There is no live PowerCenter system to connect to: extraction works entirely from the files you provide.

Run crawler​

Choose to run the crawler once immediately or schedule it to run at a recurring interval.

  1. Click Run to run once, or Schedule & Run to run hourly, daily, weekly, or monthly.
  2. Monitor progress in the workflow execution details.
note

A scheduled run reprocesses the input the workflow is configured with. Direct uploads are static: every run reads the same archive, so re-export and re-upload your files to refresh lineage. Cloud storage imports are read again on every run: replace the export at the configured prefix or key and the next run picks it up, making scheduling useful when your export pipeline refreshes the staged files. See How do I keep lineage metadata up to date?.

After the crawler completes, navigate to any resolved source or target asset and open the Lineage tab to view the data flow extracted from PowerCenter.

See also​

What does Atlan crawl from Informatica PowerCenter: Review the lineage assets and supported source types

Need help​

If you encounter issues while crawling Informatica PowerCenter assets, see Crawl and lineage issues. For additional assistance, contact Atlan support by submitting a request.