Import & Export
CSV import errors: shifted columns or changed values?
A CSV file represents values as text. When opening or importing it, the destination application decides how to interpret columns and data types. The same file can therefore look correct in a preview but change product numbers or numeric values in the next application.
- SourceOriginal export
- Check the connectionFormat & Field rules
- DestinationDestination import
How this appears in daily operation
The ERP export is available. On import, everything suddenly appears in one column, accented characters look wrong or 00123 becomes 123. Start with an unchanged copy of the original file. A file already opened and saved again in Excel is not a reliable comparison.
Distinguish typical causes
Different delimiter
Source and destination may expect different delimiters, such as commas or semicolons. An import preview helps assess this.
Character set or text qualifier
Accented characters, quotation marks and delimiters inside a field must match the expected processing.
Automatic type conversion
Excel can remove leading zeros or alter long numbers. Product identifiers are often text even when they contain only digits.
What you can check first
Keep the original file unchanged
Work with a copy and note which application and export settings produced it.
Use a preview instead of a direct import
Check columns, character set and the first representative rows. In Excel, “From Text/CSV” allows more targeted import preparation than simply double-clicking.
Compare problematic rows
Use a normal row and one with an accented character, empty value, long identifier or leading zero. The production destination remains unchanged during this check.
When the problem runs deeper
An import that is syntactically readable may still be incorrect for the business: product numbers, units or dates may not match the destination. We therefore agree field mappings, required values and handling of rejected rows. Repeated manual corrections can then become a controlled process.
Distinguish identifiers from numbers
A product number is an identifier. Having only digits does not determine its data type. The same applies to postcodes or customer codes. Source and destination must retain the same meaning; an attractive display alone is insufficient.
Repeatable imports instead of monthly rework
For a planned process, the input format, validation result and feedback are recorded. Invalid rows need recognizable handling. A new filename or resubmission must not silently create the same order a second time.
When support makes sense
If every file needs manual correction or incorrect values are discovered only later, a defined import path makes sense. An anonymized example file and a description of the destination application are enough to start.
The right solution pathAutomate file processing and data importsSources for technical assessment
Technical content reviewed on 7 October 2026. Vendor statements refer to the products named; they do not replace checking your specific configuration.
Clarify the next step together.
Your application names and a brief description of the problem are enough to start.