Jira · Data quality · 2026-11-19
Jira Assets: preparing a CSV import without creating duplicates
A practical approach to reliable Assets imports: stable identifiers, data ownership, controlled testing and checks before scheduling.
Answer this question before automating
How can you update an inventory maintained in Excel without recreating the same equipment on every import? The first decision concerns object identity, before synchronisation frequency. A name can change, an owner can leave and a spreadsheet row can move. Agree on what consistently represents the same equipment in both the source file and the destination inventory.
For Assets Cloud, Atlassian explains that matching uses an identifier attribute; a failed match can create a new object. The technical references in this article cover Cloud. For Data Center, check the documentation for your version. The preparation method below is our proposed working approach, to be adapted to your environment.
1. Define a simple data agreement
Start with one object type and one accountable source. Write down what each column means: its business definition, expected format, whether it is mandatory and who may correct it. An “owner” field must specify whether it refers to the business owner, technical support team or physical custodian. Those roles are not interchangeable, and ambiguity will survive even a technically successful import.
Then choose an identifier that remains stable within your scope. Do not assume a serial number is always populated or unique: inspect the actual values. If several entities reuse local codes, agree on a common convention before importing. Keep identity separate from the display name so that a rename does not become a new piece of equipment in your matching logic.
2. Check the file before handing it over
We recommend a preliminary check that isolates rows with missing identifiers, repeated identifiers and values outside your agreed vocabulary. Preserve leading zeros in codes and handle unnecessary spaces explicitly. Do not automatically merge two rows merely because their names look similar. Send exceptions to the source owner with a clear explanation of what needs investigating.
Also decide which source has authority over each piece of information. A technical inventory might supply an equipment model while another register defines its business owner. Without that agreement, a successful import can replace validated information with an older value. Record your rules and retain a dated source file so that subsequent changes can be explained.
3. Test difficult cases in an isolated scope
Entirely fictional example: a pilot file contains twenty equipment records identified from EQ-001 to EQ-020. Prepare a second version in which EQ-007 has a new name, one row repeats EQ-012 and EQ-020 is missing. Define the expected outcome before testing: the rename preserves identity, the preliminary check blocks the duplicate, and the absence triggers a review rather than automatic deletion.
The Cloud mapping documentation distinguishes missing objects from empty values. Review these settings separately. For this pilot, we recommend a conservative configuration and an isolated test environment. Absence from a partial export should not be treated as evidence that equipment has been retired.
4. Accept the outcome before scheduling
After testing, compare identifiers and attributes before and after, rather than checking only the total object count. Run the same file again: your acceptance criteria should include no unexpected additional objects. Document rejected rows, authorised corrections and a tested recovery procedure. A technical success message is not a substitute for business validation of the resulting inventory.
Finally, name the person who reviews discrepancies and the person who can suspend subsequent imports. Automation becomes useful when these responsibilities and checks are repeatable. ITSMWORK supports ITSM governance, Jira Service Management and reporting. A focused assessment can establish data rules and pilot acceptance criteria before a wider implementation is considered.