Import or paste UTF-8 CSV, confirm structure settings, configure field rules, apply and review differences. The download button is only enabled if the structure and type checks pass, the results are not expired, and formula risks are explicitly addressed.
Parsing and cleaning methods
The parser distinguishes between field start, unquoted field, quoted field, and closed quotes, and handles delimiters, double quotes, embedded newlines, and CRLF. Cleaning performs retain, rename, trim, null, type and deduplication in column order.
JSON accepts only arrays of objects whose values are text, numbers, booleans or null; nested values are not supported. Numbers retain their original notation, and all cells become text. null, missing fields and empty strings all become empty cells, so original types cannot be recovered. CSV output always uses commas and CRLF; all values in JSON output are strings.
Verifiable example
The tutorial example contains two lines of data, leading and trailing spaces, and a =2+3 formula risk. After applying the default rules, you should get two rows and three columns, two trim modifications, and keep download blocking until the formula policy is selected.