How the Import Error Review Sheet Works
AnyTool reviews a CSV row by row entirely in the browser. The file is parsed with PapaParse and rectangularised to the header width, then each column's dominant type is inferred from its non-empty cells. Every row is scanned for problems: a cell that does not match its column type is an error (emails checked structurally, dates validated with a deterministic UTC parser rather than the browser's locale guesser, numbers parsed locale-aware), while a missing value in an otherwise-populated column, stray whitespace, or an over-long value is a warning. A row with any error is not clean; warning-only rows are counted separately. The result is a row-level error sheet you can filter by severity and export, plus a clean-rows export. Nothing is uploaded.
- Runs client-side — the file is never uploaded
- Column-type inference drives type-aware row checks
- UTC date validation, structural email checks, locale-aware numbers
- Errors vs warnings; clean rows tracked separately
- Export the error sheet or just the clean rows
- No CDN, no backend, no row cap
