Filter CSV rows — without uploading your file.

Match ALL (AND): every rule must be true for a row to pass. Example: status = active AND revenue > 1000 keeps only rows where both are true. Match ANY (OR): at least one rule must be true. Example: country = USA OR country = Canada keeps rows from either country.

Not with the UI's simple toggle. For complex mixed logic, use Regex mode — a pattern like ^(active|pending)$ in the status column is effectively "status = active OR status = pending" combined with any AND rules outside it.

Keeps rows where the text is NOT anywhere in the cell. Example: "email does not contain gmail.com" keeps every row whose email column doesn't have that substring. Case-insensitive by default.

If the cell looks like a number (parses cleanly), it's compared as a number. Otherwise, the operator returns false. So a "price" column with "1200" vs "800" compares numerically; a free-text column with "high" vs "low" won't match any > / < rule.

Yes — "is empty" matches rows where that column is blank or only whitespace. "is not empty" matches rows where the column has any real value. Useful for catching incomplete records.

JavaScript regex (ECMAScript). Pattern flags: g for global, i added automatically unless "case sensitive" is ticked. Typical patterns: \d+, ^ABC, (apple|banana). If your pattern is invalid, you'll see an error rather than a silent no-match.

Yes. Click "Preview" and you'll see the count before committing. Saves you from downloading a file that filtered to zero rows by accident.

Not in one pass. But you can run the filter twice with opposite "Action" settings — once to keep matches, once to remove them. That gives you two output files (inside + outside the filter).

Save settings stores the output format and match-logic (AND/OR). Individual rules aren't saved because they reference specific column names that might not exist in your next file.

Yes. Output rows are the original rows, unchanged. Numbers stay numbers, leading zeros stay, dates stay in their source format. The tool only selects rows — it doesn't transform their values.

Yes — the preview updates live as you type a rule. Every time you change a column, operator, or value, the preview table and the match counter above it refresh automatically. You can see the first 10 matching rows plus the total count before committing to a download, so there's no trial-and-error.

If your dates are in a format like 2024-03-15 (ISO) you can use two rules: date column greater than or equal 2024-01-01 AND date column less than 2025-01-01. Make sure logic is set to Match ALL (AND). For other date formats, the Regex operator works — e.g. ^2024- catches every row starting with "2024-".