A PRACTICAL CHECKLIST

Clean the file.
Keep its meaning.

A tidy spreadsheet can still be wrong. Good cleanup preserves what each row represents and makes every transformation reviewable.

1. Keep the original

Work on a copy. Record row and column counts before you start. Export one Excel sheet as CSV UTF-8, then confirm whether its separator is a comma, semicolon or tab.

2. Treat identifiers as text

00102 may be a different ID from 102. Our tool keeps strings intact, but spreadsheet software can convert them at import. Set identifier columns to Text when importing. This also matters for long account numbers and SKUs.

3. Check what a duplicate means

Two matching rows could be accidental duplicates or two genuine transactions. Define the unique event or order key with the file owner. Exact-match counting is evidence for review, not automatic permission to delete.

4. A blank price is not zero

Keep unknown values blank or obtain them from an agreed authoritative source. Do not invent prices, dates or currencies just to eliminate empty cells.

5. Fix structure before transforming

Commas inside a field need valid CSV quoting: "notebook, pen" is one cell. Rows with different column counts require a structural decision. The tool will not guess which field shifted.

6. Read the change log

Review source row numbers, removed rows, trimmed cells and before/after counts. The downloaded log records all changes; the preview shows the first 20. More changes do not necessarily mean better data.

7. Be careful with spreadsheet formulas

Some text values beginning with characters such as = can be interpreted as formulas by spreadsheet software. The audit flags candidates, including some harmless negative numbers. Values are not rewritten. Review untrusted data and import as text where appropriate.