How to find files with names having a long string after the first dot

findregular expression

My Samsung Galaxy S doesn't like files where the name contains a long string after the first dot. I guess some part of the software has a too short buffer for the file extension. How can I find all files on my Micro-SD-Card mounted at /media/sd where the filename has more than five characters after the first dot?

Reading the find manpage, I thought

find . -regex '.*\..\{5,\}'

would work, but it doesn't give any hits, so I somehow don't get my regular expression right.

Best Answer

The problem is that the default type of regex's used by find is emacs-style. Find's documentation for emacs style regexes doesn't include the \{n,\} construction--leading me to believe that it isn't supported in find's implemenation of emacs-style regular expressions. The emacs-wiki lists this as valid, however it is possible that this wasn't always the case.

I found that your regex produced output if you do this:

$ find . -regextype posix-basic -regex '.*\..\{5,\}'

Related Solutions

Shell – find utility does not recognize bracket notation

It seems that the problem is \( is interpreted by the shell first.

For example, you got a file in the current directory named:

foo(.test

Then

ls foo\(.text

and

ls "foo(.text"

both work.

So, the following works:

find . -regex ".*submit\(Web\)?\.m"

EDIT:

Single quote also works here.

Difference between single quote and double quote, see this post: single quote vs double quote in bash

Where does the pattern occur in a match by find

The pattern given to -name has to match the entire base filename. The behaviour of the -name pattern is defined as:

The primary shall evaluate as true if the basename of the current pathname matches pattern

This means it's true when the whole of the basename matches the pattern you gave. You can think of a pattern as being basically like a shell glob: you can use *, ?, and [...] patterns inside it, with the start and end of the pattern aligned with the start and end of the string.

So your command:

find ~ -name bookmarks

finds files named "bookmarks" because that is the entire filename, but:

find ~ -name bookmark

would only find files named 'bookmark', because there are no wildcard characters in the pattern.

To match files called both bookmark and bookmarks, you could use:

find ~ -name 'bookmark*'

So if you want to find

those files whose names contain bookmark regardless the position of bookmark in the filename

you can use use:

find ~ -name '*bookmark*'

to match files whose names have any number of characters, then bookmark, then any number of characters.

Best Answer

Related Solutions

Shell – find utility does not recognize bracket notation

Where does the pattern occur in a match by find

Related Question