Available for projects & agency overflow · Quick reply, from the person who does the work

My file imports and exports keep failing

A catalogue file that will not go through always produces the same feeling — "but the file is fine" — and hides three very different failures. The file can be refused, truncated, or accepted but misread. The corrective work is never the same, and two of those three cases are fixed without touching the catalogue.

Describe my issue Send a message

Three failures, three signatures

  • Processing stops after a fixed time, always at roughly the same point: a time or memory limit, not a file defect.
  • Accents become symbols, or the file is read as a single column: an encoding or separator problem.
  • The file goes through entirely but the data does not match: columns are mapped to the wrong fields.
  • The export produces an empty or one-byte file: processing failed before writing, often for the same memory reason.

Encoding and separators, two spreadsheet traps

A file saved from a European spreadsheet often uses the semicolon as a separator and a regional encoding, while most importers expect a comma and UTF-8. The result is not a clean error: the file is accepted, and rows end up in a single column, or accented characters are replaced by symbols across the whole catalogue.

The second trap comes from values that contain the separator themselves: a description holding a comma shifts every following column on that row, and that row alone. That is why an import can succeed on a thousand rows and corrupt the three containing one particular character. The remedy is to wrap values in quotes, which spreadsheets do not do consistently.

Isolating the cause in four attempts

  1. Test with ten rows

    If ten rows pass and the full file fails, the cause is a resource limit. If ten rows fail too, the cause is in the format.

  2. Test with the last ten rows

    If they pass alone but not inside the full file, that confirms the limit. If they fail alone, one specific piece of data is at fault.

  3. Open the file in a text editor

    Not in a spreadsheet, which reinterprets everything. You can see the real separator, the presence of quotes and the line endings.

  4. Check what was already written

    An interrupted import often leaves part of the rows already saved. Rerunning without cleaning creates duplicates, turning an incident into a project.

Carry on with the right page

Describe your need in one minute

A few targeted questions so I can reply with an estimate rather than another questionnaire.

outil
api
sens
frequence (facultatif)
Please provide an email or a phone number so I can get back to you.

Please provide an email or a phone number so I can get back to you.

Frequently asked questions

My file works on my machine but not on the site. Why?
Because your spreadsheet displays it according to your regional settings, while the server reads it byte by byte. Opening the file in a text editor shows what the server actually sees.
Should I split the file into several parts?
An effective and immediate solution when the cause is a time limit. It fixes nothing if the cause is the format: the parts will fail the same way.
How do I avoid duplicates after an interrupted import?
By using a stable identifier as the update key rather than creating new entries. It is the setting most often forgotten in repeated imports.
Can an import break the site?
It can make it inconsistent: products without a category, wrong prices, missing images. Reversible with a prior backup, very painful without.
Is an automatic feed better?
As soon as the exchange becomes daily, yes. A manually handled file introduces human error on every run, whereas a direct link applies the same rule every time.