CSV Formula Risk Auditor
Inspect a delimited table before opening it in a spreadsheet. The tool parses quoted fields and delimiters first, then checks the start of every decoded cell, including the first row.
Key features
- Parse comma, semicolon or tab-delimited records with quoted delimiters and line breaks
- Flag = + - @, fullwidth variants and formula markers after leading whitespace or controls
- Show record, column and original text, including possible false positives such as -10
- Download a JSON finding report or a separately requested, fully quoted CSV review copy with apostrophes on candidate cells
- Process input in browser memory without automatically changing or uploading the source file
How to use
- Select a UTF-8 CSV or TSV file, or paste the table text.
- Choose the actual delimiter and run the cell audit.
- Compare the reported record and column positions with the source table.
- Decide whether a candidate is an intended number, code or formula.
- If useful, download the report or review copy and test the file in the target spreadsheet workflow.
Use cases
- Review a customer-data export before sharing it with spreadsheet users
- Find a formula-looking cell hidden after spaces or a tab
- Check that quoted commas and quotation marks do not shift columns
- Separate expected negative numbers or @ names from suspicious input
Frequently asked questions
Does zero findings prove the CSV safe in every spreadsheet?
No. This static check covers known starting characters and selected leading whitespace and controls. Import settings, spreadsheet versions and later saves can change interpretation. Verify the target workflow.
Why is -10 reported?
A leading minus is a conservative formula-prefix candidate, but -10 may be an intended negative number. The auditor reports position and original text rather than claiming a formula executed.
Is wrapping a field in double quotes enough?
No. Quoting preserves the CSV field structure; it does not guarantee that a spreadsheet will treat a formula-like value as text. Save-and-reopen behavior may differ.
Is the review copy identical to the source data?
It retains row and column structure and leaves noncandidate cells alone, but adds one apostrophe before each candidate value. Those underlying cell strings therefore change. This mitigation is not universal across applications or repeated saves.
Does the tool repair blank or ragged rows?
No. It reports inconsistent column counts, including ambiguous internal blank rows, and does not produce a review copy from such a table. The first row is inspected like any other record.
Is my file uploaded?
No. The browser reads the file locally. The JSON report includes original candidate values and the review CSV includes the whole table, so inspect downloads before sharing.
Privacy
Selected tables are processed in browser memory and are not uploaded or automatically saved. An explicitly downloaded report can contain sensitive original cell values.
Comments & questions