Filter Lines (Keep or Remove Lines Containing)

About the Filter Lines (Keep or Remove Lines Containing)

Paste a list, a log file or any text and pick the lines you want: keep (or remove) the lines that contain a word, start with it, end with it or are exactly it — or that match a regular expression, or that are longer or shorter than a number of characters. Type several terms, one per line, and choose whether a line must match any of them or all of them.

The result shows the kept lines and the removed lines side by side with their counts, so you can check at a glance that nothing important went missing, and copy or download either list. It works like grep (and grep -v) on the command line, but needs no install and keeps your text on your device.

Common uses

• Pulling errors, a user ID or an order number out of a server or application log.
• Removing unsubscribed, bounced or internal addresses from a mailing list before an import.
• Keeping only the URLs of one domain or folder from a sitemap or crawl export.
• Finding lines that are too long for a subtitle, an SMS or a database column.

How the matching works

• Contains, starts with, ends with and is exactly compare plain text: a dot is a dot and a bracket is a bracket. Ignore case uses full Unicode case folding, so STRASSE finds Straße, and an accented letter typed in either of its two Unicode forms matches the other.
• Whole words only (for contains) needs a non-letter, non-digit character or the line edge on both sides: cat finds “the cat sat” and “cat-like”, not “category” or “bobcat”.
• Ignore spaces at line ends lets starts with, ends with, is exactly and the length filters look past indentation and trailing spaces — and a stray space at that end of a term is ignored too. The lines you get back are never changed.
• Any term / All terms: with several terms, Any keeps a line that matches at least one; All needs every one of them (for example, lines that contain both “invoice” and “overdue”).
• Line length counts the characters you see — an emoji or an accented letter counts once, even when it is stored as several code points.

Regular expressions

Each line of the terms box is one JavaScript regular expression, tested against each whole line: ^ and $ mean the start and end of the line, \d a digit, \b a word boundary, (a|b) either. Matching runs in a background worker with a 3-second limit, so a pattern that would take forever (catastrophic backtracking, such as (a+)+$) is stopped instead of freezing the page. To try a pattern on sample text first, use the Regex Tester.

Command-line equivalents

Keep · contains is grep -F, Remove is grep -v, Ignore case is -i, Whole words only is -w, matches regex is grep -E (with JavaScript syntax), and Add original line numbers is -n.

Limitations

• Lines are filtered one at a time; a term never matches across a line break.
• Regular expressions use JavaScript syntax. Lookbehind, named groups and Unicode property escapes such as \p{L} work in current browsers; POSIX classes such as [[:digit:]] do not.
• Whole-word matching treats letters, digits and the underscore as word characters; it does not know about compound words in languages written without spaces.
• Files up to 50 MB can be opened; very large results are previewed in part, while Copy and Download always include everything.

Privacy

Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.