Follow the file into the workspace
Upload a DataStage XML export and follow upload progress separately from parsing activity. Resumable uploads let you select the original file and continue from the saved position after an interruption.
Import activity records the filename, project, queue and completion times, parsing duration and status. A cancelled or incomplete import can retain earlier complete jobs; inspect the recorded result before deciding whether to upload again.
Account for the jobs in the export
The diagnostic summary separates jobs encountered, jobs imported and jobs not imported. It also records whether the complete XML document was read. Those distinctions help you tell an incomplete file from a completed import that contains issues.
Use job capture counts to inspect the structures recorded for a job. A successful upload establishes that the file arrived; the import report explains what reached the Catalog.
Separate source capture from conversion
Review import errors, source capture findings, job capture counts and conversion issues as separate evidence categories. Search by job, stage, issue code or explanation, then filter conversion findings for Microsoft Fabric, Databricks or Snowflake.
A source definition can be captured while still containing a stage or expression that blocks conversion for the chosen target. Keeping those findings separate helps the export owner and migration engineer identify their next actions.
Carry diagnostics into remediation
Expand a finding to read its explanation, affected job or stage, issue code and target applicability. Download the diagnostic report as JSON to retain the recorded import evidence with an engineering investigation.
For example, first confirm that all expected jobs were read. Next filter for the intended destination and investigate its blockers. Then inspect the source-capture findings with the DataStage owner before generating a representative target workload.
Use import evidence for the right decision
Import diagnostics establish what was captured and which issues were recorded. They do not establish that generated code will produce equivalent results. Resolve conversion blockers, review source gaps and run the target against controlled inputs as separate migration steps.
Start with a DataStage XML job export. Runtime schedules, external scripts, credentials and production data require their own inventory alongside the design metadata.