I have a fair few files that have been extracted from a zip created in a strange way.
the files have come out of the tar.gz in windows file structure format
Example:
jpg_250\MI\00\00\00\19\MI0000001900.jpg
versus
jpg_250/MI/00/00/00/19/MI0000001900.jpg
The former is seen as a single file by linux.
I've been playing around with awk and sed to delimit the filename by backslash, and create the dir开发者_运维知识库ectories in question in the correct structure, and finally rename the file to the MI**.jpg and move it into the correct newly created end directory.
Is awk and sed the way to go here? I have awk exploding the filename into the 'directories' I need but I'm having trouble getting the directories actually created. I assume I would need sed at the end to rename the file into the MI**.jpg format.
Many thanks for any help.
Something like this?
$ ls
a\b\c
a\b\d
$ for i in *; do
F=$(echo $i | sed 's,\\,/,g')
D=$(dirname $F)
echo mkdir -p ${D}
echo cp "${i}" "${F}"
done
mkdir -p a/b
cp a\b\c a/b/c
mkdir -p a/b
cp a\b\d a/b/d
You could do this with Perl like this.
# assume that the script was called like this:
# myscript file1 file2 etc...
# then all the files are in @ARGV
foreach $orig (@ARGV) {
$orig_fixed = $orig;
# convert the \ to /
$orig_fixed =~ s!\\!/!g;
#split the filename into directory and filename
($dir, $filename) = ($orig_fixed =~ m!^(.*)/([^/]*)$!);
# create the directory if it doesn't exist
if (!-e $dir) {
`mkdir -p $dir`; # -p means create the full path
}
# now move the file
`mv '$orig' $dir/$filename`;
}
精彩评论