Data import: validating records and understanding partial results

Import is a way to create or update many records consistently. It is valuable when the source data is prepared carefully, but volume amplifies mistakes. A wrong reference or duplicated identifier can affect many records at once.

Before importing, define what each row represents, which fields identify an existing record and whether the goal is to add new data or update existing data. Those are different operations with different consequences.

DNA's import workflow

DNA's Data Import page supports master-specific templates, parsing, validation, reference lookups and row review. It distinguishes Add and Update modes. The Add flow requests duplicate skipping without updating existing records, while Update uses a separate bulk-update operation.

The bulk service processes rows marked valid in batches and accumulates succeeded, failed and skipped results. A finished batch process is therefore not evidence that every row succeeded. Reconcile the outcome to the source population.

Reference-data validation has a condition

DNA attempts to load lookup lists before validating related values. If a lookup fails, it reports that the affected values were not checked against the master. A row marked valid under that condition should not be interpreted as fully verified reference data.

For example, a supplier category can only be checked against the available category definitions when those definitions loaded successfully. An empty dropdown caused by failure is different from a valid list containing no allowable values.

Recover deliberately

Review failed and skipped rows before repeating an import. A blind repeat can create confusion about what already succeeded, particularly when some batches completed before a later failure.

Keep the source file, mode, scope and result summary identifiable. Validate representative saved records and the total outcome. Import success is a reconciliation exercise, not merely reaching the end of a progress bar.