Duplicate asset records created after bulk import in asset lifecycle module, unique constraint not enforced

PLM Aras Innovator
, , , , , , , ,
1

We’re experiencing a critical issue with our asset lifecycle module after performing a bulk import of approximately 5,000 asset records from our legacy system. The import completed without errors according to the import tool logs, but we’re now seeing duplicate asset records in the database - some assets appear 2-3 times with different IDs but identical serial numbers and descriptions.

Our import mapping validates the serial_number field as the unique identifier:

<mapping>
  <field source="SerialNo" target="serial_number" unique="true"/>
  <field source="AssetDesc" target="description"/>
</mapping>

We have a database unique constraint on the serial_number column, but somehow duplicates still got through. I’m concerned about data integrity and need to understand how this happened before we can safely clean up the duplicates. Has anyone dealt with import-related duplication issues in asset lifecycle management?

2

For large imports, you should always implement pre-import deduplication in your ETL process. Create a staging table in the database, load your source data there first, then run SQL queries to identify and resolve duplicates before the actual Aras import. This gives you full control over merge logic and prevents the exact scenario you’re experiencing. The import tool is great for getting data in, but not for complex validation.

3

Good point. I checked the Asset ItemType and the serial_number property doesn’t have is_unique set. That explains why the database constraint wasn’t preventing duplicates during import. But now I’m wondering about the pre-import deduplication strategy. Should we be running a validation script on the source data before importing, or is there a way to configure the import tool to check for duplicates before creating new records?

4

I’d add that you need to consider case sensitivity and whitespace in your serial numbers. Even with unique constraints, ‘SN-001’ and 'SN-001 ’ (with trailing space) would be treated as different values. Your source data cleanup should include trimming and standardizing the format of unique identifiers before import.