I want to delete every regular file whose name ends with .mp4. This command worked, although the trailing wildcard seems unnecessary: find -type f -name '*.mp4*' -delete. I first tried find -type f -name '.mp4$' -delete, expecting $ to mean "the end of the filename," but it matched nothing. What is the correct way to express an ending match with find, and why did the original pattern fail?
3 Answers
The -name predicate uses shell-style glob patterns, not regular expressions. In a glob, * means “any sequence of characters,” while $ has no special meaning, so '.mp4$' means a filename that literally ends in .mp4$. To match files ending in .mp4, use: find -type f -name '*.mp4' -delete. Keep the pattern quoted so the shell does not expand it first.
Glob patterns are effectively matched against the whole basename. That means '*.mp4' already means “anything followed by .mp4,” so it is the glob equivalent of the regular expression '.mp4$'. By contrast, '*.mp4*' also matches names with extra characters afterward, such as video.mp4.backup.
If you specifically want a regular expression, GNU find provides -regex. Since it matches the complete path, a typical command would be: find -type f -regextype posix-extended -regex '.*.mp4$' -delete. The dot must be escaped because an unescaped dot in a regular expression means any character. For ordinary filename suffix matching, though, -name '*.mp4' is simpler and more portable.
The exact regular-expression syntax depends on the find implementation and its selected regex type, but the important distinction is that -regex is needed for regex operators such as $; -name only accepts glob syntax.