CSV export help · ERP and supplier files
Fix CSV encoding and mixed line endings
An old ERP export can use a different encoding, semicolons or several kinds of line endings. Keep the original file, then check the text before converting it.
Start with the original export
- Keep an untouched copy. If possible, export again from the source as CSV UTF-8.
- Open the free CSV fixer and select your file. It runs in your browser, without uploading the file, up to 15 MiB.
- Check the detected encoding, source delimiter and preview. Look at accented names and references such as
00123. If they are wrong, stop and return to the source export. - Choose the output delimiter and line endings. For our paid Shopify preparation, choose comma and LF, then download the UTF-8 CSV.
Shopify requires UTF-8 CSV files and comma-separated headers. Shopify’s CSV guidance explains these requirements. Changing the encoding alone does not supply missing product columns.
Which encodings can the fixer read?
- UTF-8, with or without a byte order mark (BOM).
- UTF-16 LE or BE, when the file starts with its BOM.
- Windows-1252 as a fallback for some files that fail the UTF-8 check.
The fallback is a heuristic, not a universal encoding detector. Some mixed or invalid encodings are refused; other code pages can decode into incorrect text. A correct-looking encoding label is not proof that every character is right. For a different ERP code page, request a UTF-8 export from the source.
Can conversion restore é, replacement characters or lost accents?
Only if the original bytes still contain the intended characters in a supported encoding. Text that was decoded incorrectly and then saved again, or characters replaced by � or ?, cannot reliably be restored by re-encoding. The fixer does not guess the missing text. Return to the original export.
Mixed CR, LF and CRLF: records versus cells
The fixer accepts CR, LF and CRLF between records, even when they are mixed. It rewrites those separators as your chosen LF or CRLF. A line break inside a correctly quoted cell stays inside that cell and is preserved.
Fields containing commas, quotes or line breaks need correct quoting; quotes inside quoted fields are doubled. RFC 4180, section 2, documents these CSV conventions.
SKU,Title,Description
00123,"Mug, summer","First line
Second line"
This fictional excerpt has one product record after its header, even though its description spans two physical lines. It is not a complete Shopify import file.
What conversion does not repair
- Unclosed or misplaced quotes: conversion is refused.
- Different column counts: the fixer warns and preserves the rows; correct the source before using them.
- Lost SKU zeros:
00123is kept when present, but123is not padded by guessing. Read the SKU and quoting guide. - Missing prices, variants or product mapping: these are separate issues.
Before a paid preparation order
Our €20 service accepts up to 100 source rows and 1 MiB, for new simple products with supplied EUR selling prices. Upload a valid UTF-8 CSV with commas; CR-only record endings are refused, so choose LF in the fixer first. These are our service limits, distinct from the free fixer’s limits.
The paid form uploads the source privately and asks you to confirm the mapping and scope. Your complete ZIP is checked and prepared before payment. We do not import into your store, and an actual Shopify store import has not been validated. The encoding examples were tested on synthetic files, not on a particular ERP’s export.