I have a list of files, in different folders, named like
aaaaaa_bb_cccc_ddddd_ee.jpg In some of them, I need to remove everything up to the second _, so they become
cccc_ddddd_ee.jpg In others (already in a separate folder), I need to remove everything up to the third _
ddddd_ee.jpg I know commands such as rename, which use some REGEX, but I don't know the exact expression for this case. How can I do that on Linux terminal?
With the Perl Rename utility,
$ ls aaaaaa_bb_cccc_ddddd_ee.jpg $ rename -n 's/([^_]*_){2}//' * rename(aaaaaa_bb_cccc_ddddd_ee.jpg, cccc_ddddd_ee.jpg) $ rename -n 's/([^_]*_){3}//' * rename(aaaaaa_bb_cccc_ddddd_ee.jpg, ddddd_ee.jpg) The -n flag tells Rename to output what it would do. If you remove -n, the modifications will be applied.
The * will expand to every file in the current directory. In the example above, there was only one file, but Rename can operate on multiple files in a single run, in which case * is faster instead of a shell loop.
s/([^_]*_){2}// means: substitute every occurrence of the ([^_]*_){2} regular expression in filenames by nothing. The regex means: any number of non-underscores [^_]* followed by a underscore _ repeated twice.
To eliminate everything up to the second _
rename -n 's/[^_]*[_][^_]*[_]//' * To eliminate everything up to the third _
rename -n 's/[^_]*[_][^_]*[_][^_]*[_]//' * Remove the -n to effectively change the filenames, and not just test.
* as globs (highly unlikely in this case, though, you would have to have a very weirdly named file for that to happen.).