Recover a failed or partial import
Read import outcomes, correct only the missing records, and avoid duplicating successful rows.
An import can create some records while leaving others behind. Read the final result before uploading the same file again.
Identify the outcome
Section titled “Identify the outcome”| Message or result | What it means | Next step |
|---|---|---|
| Some rows will not be imported before submission | The selected valid rows can proceed; the listed rows still have errors. | Cancel to fix them first, or deliberately import the valid subset and keep a list of the remainder. |
| Import finished with skipped rows | Some records were created or skipped; others were not imported. | Compare Imported, duplicate/skipped, and Not imported counts. Retry only confirmed missing records. |
| Nothing was imported with an all-or-nothing rollback message | That batch was rolled back. | Fix the reported problems and review the corrected batch again. |
| Imported — but some columns were not | Rows were created, but unsupported column values were discarded. | Inspect those fields in the saved records. Do not reimport the whole file merely to add an unsupported column. |
| A timeout or interrupted connection with no final result | The write outcome is unknown. | Inspect the destination list and import activity before retrying. A missing success message does not prove nothing was created. |
The all-or-nothing statement applies only when that mode and rollback outcome are explicitly reported. It is not a blanket guarantee for every import route.
Correct the missing records
Section titled “Correct the missing records”- Keep the original file and record the import time, resource type, and outcome counts.
- Choose Download error report (CSV) when offered. For manual journals, use Download not imported (Excel) where available.
- Read each row’s field and error reason. Fix the original source data: missing contacts/accounts, invalid dates, duplicate references, unbalanced journals, or other stated errors.
- Prepare a separate file containing only records confirmed not to have been created. A CSV error report is a diagnostic report, not necessarily an upload-ready copy of your data.
- For a multi-line invoice, bill, or journal, retain the complete document’s lines together. Do not retry an isolated error line as a new standalone document.
- Review the new mapping and preview, submit the corrected batch, then compare the destination records and totals with your source.
If the number not imported is larger than the itemized error list, the displayed list is incomplete. Reconcile against the original file and saved records; do not treat the short list as the complete retry set.
Finish the reconciliation
Section titled “Finish the reconciliation”For a file of 20 records with 17 created and 3 not imported, verify those 17 in the destination before retrying the remaining 3. Deduplication behavior varies by resource and reference; do not assume uploading all 20 again is harmless.
Check counts and key amounts/dates. For financial documents, compare reports; for bank activity, review duplicate date ranges and match existing entries.
Related walkthrough
Section titled “Related walkthrough”Compare the bank import preview example with your own Results. A preview marked Ready is not proof that rows were saved; verify the destination after submission.