CSV Import Errors: How to Repair a Rejected File

Start with the error, not a bulk cleanup

Before changing data, keep an untouched copy of the exported file and record the receiving application's expected column names and formats. A parser that accepts a file is not proof that its IDs and relationships are intact.

Five checks before re-importing

1. Headers: compare exact names and required fields against the target import template; avoid blank or duplicate headings.
2. Row width: count values after parsing quoted commas, escaped quotation marks and multiline fields.
3. Identifiers: keep product codes, postal codes and customer IDs as text so leading zeros survive.
4. Dates and numbers: apply the destination's expected format without guessing locale or converting blanks into zero.
5. Conservation: compare original and candidate row IDs, duplicate counts and unrelated columns before submitting.

Example: why totals are not enough

record_id,amount
A101,10.00
A102,20.00

If a revised CSV changes those two amounts to 15.00 each, the total still equals 30.00—but both records changed. An import check should report the row-level discrepancy even when the grand total balances.

When to request a repair

If the application refuses the file, send the error text and the destination's import requirements first. Do not email private customer data to an unverified address. Our browser-based demonstration can compare small local CSV samples without uploading them.