Find files filtered by multiple extensions

Question

What is the correct syntax for:

find . -type f -name \*.\(shtml\|css\)

This works, but is inelegant:

find . -type f -name \*.shtml > f.txt && find . -type f -name \*.css >> f.txt

How to do the same, but in fewer keystrokes?

Solution 1

You can combine different search expressions with the logical operators -or or -and, so your case can be written as

find . -type f \( -name "*.shtml" -or -name "*.css" \)

This also show that you do not need to escape special shell characters when you use quotes.

Edit

Since -or has lower precedence than the implied -and between -type and the first -name put name part into parentheses as suggested by Chris.

Solution 2

Here is one way to do your first version:

find -type f -regex ".*/.*\.\(shtml\|css\)"

Solution 3

You need to parenthesize to only include files:

find . -type f \( -name "*.shtml" -o -name "*.css" \) -print

Bonus: this is POSIX-compliant syntax.

Solution 4

I often find myself ending up using egrep, or longer pipes, or perl for even more complex filters:

find . -type f | egrep '\.(shtml|css)$'
find . -type f | perl -lne '/\.shtml|\.css|page\d+\.html$/ and print'

It may be somewhat less efficient but that isn't usually a concern, and for more complex stuff it's usually easier to construct and modify.

The standard caveat applies about not using this for files with weird filenames (e.g. containing newlines).

View more solutions

Comments

[Teddy]

That will also print directories named "*.css".

[Chris Johnsen]

Hmm, the parentheses in your updated version are a bit misplaced. The individual parentheses need to end up as separate parameters to find, so they need spaces around them (` ".css") ` results in a single string value; it is the same as (e.g.) ` '.css)' ). Second, the parentheses need to go around whole ‘primaries’ (the open parenthesis needs to be before -name`, not between it and its ‘operand’).

[Cristik]

+1 for a clean and modular solution, the performance bottlenecks usually occur while processing the files resulted from the search results.