How to use this tool
Both files must have the same unique column names. Enter the exact key header, such as order_id. Column order can differ. Comparisons preserve spaces and are case-sensitive. Different text representations of an amount count as changes.
A repeated key in either file is marked for review. The tool never silently merges rows or chooses a winning record. Changed-field details show the before and after values. Expand repeated groups to inspect their records, and use the download for every source record in each group.
Check the built-in sample
The sample produces two unchanged keys, one changed key, one only in A, one only in B, and one repeated-key group. Six distinct keys account for six rows in A and five in B.
Limits and the downloaded report
The parser supports comma, semicolon, and tab delimiters, quoted separators, escaped quotes, and multiline fields. It reads UTF-8 text, with an optional byte-order mark. An unclosed quote or an invalid field boundary stops processing. Large files should be handled in an appropriate local spreadsheet or data application.
Data remains in memory until you clear or leave the page. The tool does not save it to browser storage. Reports retain unresolved findings. Formula-like text is prefixed with an apostrophe in downloaded review reports; this is a deliberate report-only change. CSV does not define text column types, so import identifiers as Text.
Follow the complete example
Compare two order exports without silently merging records ↗