Inspect how CSV columns might be interpreted
CSV stores cell text rather than declared types. This inspector scans each column and reports a candidate category based on its nonblank cells. Paste a table and choose Inspect column types. The report includes blank counts, leading-zero integer strings and integer strings outside JavaScript's safe-integer range. No cell is converted or rewritten.
Example: numeric-looking IDs are still identifiers
Load the example with IDs 001 and 002, scores 10 and 12.5, true/false active values and date-like strings. The ID column is classified as integer text and flagged for leading zeroes in both cells. Scores are mixed integer and decimal syntax, so the candidate is numeric text. Active is boolean text; the dates use the supported ISO date-only format. These labels are observations of spelling, not permission to change an ID into a number.
Classification rules
Whitespace is trimmed for inspection only. Empty or whitespace-only cells count as blank and do not determine the candidate. Case-insensitive true and false are boolean text. Finite decimal and exponent notation is numeric text; integer spelling is reported separately. Exact YYYY-MM-DD strings with a real calendar date are date text. Unsupported formats become text. More than one kind becomes mixed/text, except integer-plus-decimal columns which become numeric text.
The tool does not infer local date formats, time zones, currencies, percentages, telephone numbers, semantic IDs or database schemas. A date-like value can still be a product code. An empty column has an empty candidate, not a guessed data type. A nonzero numeric spelling that underflows to zero, such as 1e-400, is classified as text rather than a usable numeric value. Candidate detection cannot guarantee compatibility with a particular spreadsheet or database importer.
Precision and diagnostics
Leading-zero flags identify integer strings such as 001 that would lose spelling if coerced. Unsafe-integer flags identify integer strings that JavaScript cannot represent exactly as safe integers. Large values may also be reported as text if they exceed finite numeric range. The JSON export includes per-category cell counts for deeper review; it is a diagnostic report, not a replacement CSV.
Input supports 200,000 characters, 200 unique nonempty headers and 10,000 equal-width data rows, with quoted commas and line breaks. Processing stays in the browser, and edits clear previous reports. Use CSV Statistics Calculator for an explicitly selected numeric column, or CSV Empty Column Remover to remove fields after checking that they are truly unused.





