Troubleshooting Clara CSV import
Solve the most common problems when importing mileage receipts in CSV format to Clara.
Why imports fail (and why it's rarely Clara's fault)
When a mileage CSV import fails in the Clara dashboard, the first reaction is usually to blame the tool. In almost every case, though, the failure comes from a detail in the file: altered encoding, swapped separators, reordered columns, or outdated cost centers. This guide solves the most common problems and shows how to prevent each one. Spending 15 minutes understanding these categories avoids hours of rework throughout the year and eliminates the bulk of support tickets opened by finance.
UTF-8 encoding errors
The typical symptom is Clara accepting the file but showing broken accents (characters like "ã" instead of "ã"). The cause is almost always opening the CSV in Excel to "take a look" before uploading, because Excel reopens it as Windows-1252 or ANSI and loses the UTF-8 marking. The fix is simple: never open the CSV exported by Quilometragem before uploading it to Clara. If you need to inspect the content, use a plain-text editor (VS Code, Notepad++, Sublime) that preserves the encoding. If the file is already altered, generate a new one in Quilometragem.
Swapped separators (comma vs semicolon)
Clara accepts CSV with comma or semicolon as separator, but it requires consistency: the entire file must use the same one. When someone edits a line in Excel and saves it as CSV on a Windows system with localized regional settings, the separator can switch to semicolon while the rest of the file still uses comma. Clara will stop the import at the first inconsistent line. The rule is the same as before: don't edit the generated file. If you need to add an expense manually, do it from the Clara dashboard, not in the CSV.
Missing or reordered columns
Clara expects a specific column sequence, and changing the order or removing any one breaks the import. Quilometragem always exports with the right sequence, but spreadsheets downloaded from older Clara templates may have a different order than the current version. When the error message mentions "missing required field" or "unrecognized column," re-export from Quilometragem instead of trying to manually reorder. The template version is automatically kept in sync.
Decimal separator and date format
Monetary values in Latin America traditionally use a decimal comma, but the international CSV standard asks for a dot. Quilometragem always generates with a decimal dot because that's what Clara expects. When someone opens the file in localized Excel, the program converts to comma automatically, and the saved CSV becomes invalid. The same applies to dates: the standard is ISO (YYYY-MM-DD), and Excel may reformat to DD/MM/YYYY, breaking recognition. A plain-text editor avoids both problems at once.
**Duplicate detection and unknown cost centers**
Clara identifies duplicates by the receipt's external identifier. If you re-export the same batch without changes, Clara warns that everything already exists. If you re-export with some new receipts mixed in, it only imports the new ones. That behavior is desirable but confusing when the user expected to re-import everything. To force reimport, delete the old entries in the Clara dashboard before uploading the new file. The unknown cost-center error appears when the project was renamed or deactivated between CSV generation and import.
The preventive workflow recommended by Quilometragem
The safest way to avoid problems is to follow a standard flow: generate the CSV, verify totals in the Quilometragem dashboard, download, upload directly to Clara without opening, validate the preview, and confirm. This flow eliminates more than 90% of support tickets. For high-volume companies, we recommend a single owner for the monthly upload, trained on this flow, instead of distributing the task across several employees. Documenting the procedure in a short runbook also helps continuity.
Next steps when the problem persists
If the error persists even with the right flow, three actions solve almost every remaining case. First: generate the CSV in a smaller batch (one employee at a time) to isolate the problematic entry. Second: compare the exported CSV against Clara's current reference template to confirm versions. Third: open a simultaneous ticket with Quilometragem support (which sees the generator) and Clara support (which sees the importer). The combination of these three actions closes any edge case within hours.