WooCommerce CSV encoding and delimiter errors: repair shifted columns

Distinguish three different problems

Encoding determines how characters are stored. The delimiter separates fields. Quoting keeps delimiters and line breaks inside a field. Changing one setting does not automatically repair the other two. Keep the original file so each experiment can be compared with the source.

Inspect the header and a representative data record in a plain-text editor. Note whether fields are separated by commas, semicolons or tabs. A semicolon-separated export can be valid tabular data, but the importer must be configured for its actual delimiter. Do not rename a file extension and expect its contents to change.

Export with an explicit preview

Use your spreadsheet’s text import or export dialog and select the intended character encoding and separator. Preview names containing accents and a description containing punctuation. LibreOffice’s CSV documentation explains the settings; WooCommerce’s core importer expects UTF-8 data.

Let a CSV writer quote cells containing commas or newlines. Do not use global search-and-replace on separators: a comma inside a product description is content. Likewise, counting text lines is not a reliable product count when descriptions contain quoted line breaks. Count parsed records instead.

Check a troublesome record end to end

The example has one product name containing a comma. The quotation marks keep it in a single Name field. Reopen your exported file with explicit import settings and check that the SKU, name and price still occupy their intended columns.

Use a test set with an accented name, a long description and a quoted field. Import those records on staging, then compare the editor and a new export. If strange replacement characters are already saved in the source, merely resaving as UTF-8 may not recover the original text. Return to an undamaged source or ask the catalog owner for the correct spelling.

Inspect the CSV structure

Valid quoting example: the comma inside the product name stays inside that cell.

Type,SKU,Name,Regular price
simple,00041,"Mug, blue",12.50

Check the result

• Every parsed record has the expected field structure.
• Names and descriptions retain their characters and punctuation.
• The mapping screen shows distinct, correctly named columns.