Amazon Inventory Ledger CSV and Excel Import: Preserve Dates, IDs and Quantities
Prepare Inventory Ledger files for review without corrupting identifiers, date formats, quantity signs or report coverage in CSV and Excel.
Keep the original export, import identifiers as text, verify the separator and date format, preserve quantity signs and avoid combining overlapping exports blindly. For FluxPilot, use a supported report format within the 20 MB limit.
Save an untouched copy before editing
Spreadsheet software can change data when it opens or saves a file. An identifier may lose a leading zero, a date may be interpreted using the wrong locale, or a numeric value may be converted into scientific notation. Keep the original export so you can distinguish a source issue from an editing issue.
Create a separate working copy and note every transformation. If you change a separator, convert a date or add a review column, another reviewer should be able to reproduce it. A clean import means preserving meaning, not simply making the file look tidy.
1. Confirm the delimiter and text encoding
If a report opens as one long column, inspect the separator before editing the contents. A file extension alone does not guarantee that the values are comma-separated. Use the spreadsheet import dialog to choose the separator that matches the file, then verify several rows in the preview.
Check product text with accents and punctuation. Keep quoted values intact when a text field contains the delimiter. Do not run a global replace on commas to fix a separator problem; that can alter titles, decimal numbers and legitimate field contents.
For a working CSV, use a consistent encoding such as UTF-8 and verify the saved result. Preserve the unchanged source file even when the converted copy looks correct.
2. Import identifiers as text
Set inventory identifiers and source references to text before completing the import. Treat them as labels, not quantities. Check a few values against the original export, especially values with leading zeros, long digit strings or punctuation.
Do not use product titles as a replacement for identifiers. If a working copy has already altered an ID, return to the source rather than trying to reconstruct it from the visible cell. A wrong identifier can create a false unmatched incident even when the compensation evidence is present.
| Data | Working treatment | Example problem |
|---|---|---|
| Inventory IDs and references | Text | Leading zeros removed |
| Event dates | Explicit, verified date interpretation | 03/04 read as wrong month |
| Movement quantities | Numeric; retain source sign | Negative values made positive |
| Currency | Separate explicit field | USD and EUR mixed |
| Product titles | Readable labels | Used as the only match key |
3. Verify dates and quantity signs
A date such as 03/04/2026 can mean March 4 or April 3. Confirm the export's interpretation before converting it to a consistent format. If the source includes times or a timezone, preserve them where relevant and note any deliberate conversion.
Keep the sign of each source movement. If you want a positive affected-units column for review, add it separately with an explanation. Converting every negative quantity to positive can destroy the distinction between an outward movement and a later correction.
Also distinguish a blank numeric cell from a real zero. A blank can mean unknown or absent information. Replacing it with zero without a documented reason can hide missing quantity or cost data.
4. Check coverage and duplicate files
Maintain a small import register with filename, account, marketplace, period, export timestamp and whether the file replaces or supplements an earlier one. Two files can contain the same source events even when their filenames differ.
Do not remove rows solely because product and quantity match. Two real incidents can share those values. Look for a stable source relationship and compare date, category and reference. When in doubt, retain the row and mark the duplication question for review.
Start with a small set of known rows to check the result. Compare the displayed identifiers, dates and quantities with the source before trusting a large import.
5. Use the formats supported by FluxPilot
The current FluxPilot Recover upload interface accepts CSV, TSV, TXT, XLSX, XLS and text-based PDF ledger files up to 20 MB. Its optional cost and reimbursement uploads accept CSV, TSV, TXT, XLSX and XLS. A scanned image is different from a text-based PDF; prefer the original structured report when available.
If the upload result is unexpectedly empty, first verify that you supplied an event-level ledger with relevant activity. Then check headers, coverage and the values in known rows. Save a corrected working copy without replacing your original source.
The reconciliation worksheet linked below is for your manual review notes. It is not an Amazon Inventory Ledger export and should not replace that source file in the report-analysis workflow.
Frequently asked questions
Can I upload a screenshot instead of the report?
Use a supported source report. The current interface describes text-based PDFs, not screenshot-based report extraction.
Should I combine all monthly exports into one file?
First verify scope and overlap. Keep an import register and ensure the same incident is not treated as new activity multiple times.
Sources and further reading
Workflows and examples are from FluxPilot. Amazon policies change; verify the current policy for your marketplace before filing a claim.
FluxPilot report workflow and supported inputs