Repair a CSV for Excel without losing IDs

Step by step

1. Open CSV Repair and select the original file, or use the included synthetic example.
2. Inspect the detected encoding, separator and line endings. If characters or columns look wrong, choose the actual input settings and inspect again.
3. Compare the imported cells with the editable preview. Apply only the corrections you need and review every warning.
4. For the basic training exercise, leave every column as Text and pad the short row. The amount -12,50 keeps its original spelling; identifiers retain their zeros.
5. Confirm the reviewed values and prepare CSV or XLSX. Download the generated file and check its import in your target application.

One column, broken accents, or missing zeros?

If the entire row appears in one column, first check whether the source uses commas, semicolons or tabs. Separators inside quoted values belong to the cell. A quoted line break belongs to that same cell too; splitting the file at every comma or newline damages such records.

UTF-8 and Windows-1252 interpret some bytes differently. Automatic detection can suggest an encoding; it cannot reliably infer the original encoding of every short file. Compare names with the source. If the file already contains replacement characters, choosing another encoding may not restore the lost characters.

A spreadsheet may interpret 0012 as a number when opening CSV directly. Import that column as text, or use an XLSX file with explicit text cells. If only 12 remains in the source, this tool does not know how many zeros were lost.

The included example

The sample contains an accented name, an embedded comma, a quoted line break, a long identifier, a negative decimal amount and a short final row. It is synthetic and contains no customer data.

Padding the last row adds an empty fourth cell. It does not invent a note. The ID 9007199254740993 must remain text: treating it as a JavaScript number would lose precision.

A reusable exercise for spreadsheet training

Download the synthetic workflow training ZIP linked below, or choose Use example in the tool. No private file is needed. The package includes CSV and list-comparison exercises with expected results and usage terms.

Basic exercise: keep every column as Text and apply only padding. Expect four data rows plus the header, with four cells per row. XLSX C2 remains the text -12,50; 0012 and 9007199254740993 also remain text. The quoted two-line note stays in D2 and the final note is empty. =1+1 in D4 stays text in XLSX. Keep CSV formula protection enabled: it exports that cell as '=1+1 and discloses the added apostrophe.

Optional extension: deliberately convert an amount

After completing the basic exercise, select only amount as Number and choose the decimal comma. Apply the conversion before exporting XLSX: C2 should now contain the number -12.5. This extension changes the amount's cell type and decimal spelling; it is separate from the package's all-text baseline. Keep the identifier columns as Text.

CSV or XLSX?

CSV stores characters, not durable spreadsheet column types. Choose the separator and encoding for the receiving application. UTF-8 with BOM helps Excel identify the encoding, but does not prevent automatic number or date conversion.

XLSX preserves explicit text cells and can carry numbers you have deliberately selected. Formula-like imported text is not executed by this tool. The CSV protection option prefixes potentially executable cells with an apostrophe; that changes the exported text and is reported. Use explicit XLSX text cells when you need to preserve the original characters without creating formulas.

What this workflow can verify

The exporter reopens the generated bytes and compares the intended cell values. This checks the file produced here; it cannot promise identical import settings across every spreadsheet version.

The workspace accepts up to 5 MiB, 10,000 rows, 256 columns and 100,000 cells. Processing runs in a cancellable local worker. This guide contains no file input: opening the workspace loads a separate document without advertising scripts. Files are not sent to the advertising provider.

Further reading

• Download synthetic workflow training ZIP — examples, expected results and usage terms
• Microsoft: import and export CSV
• Microsoft: open UTF-8 CSV correctly