How do I match exactly the $TARGET_NAME value from find /tmp -type l -exec ls -l output?
$ find /tmp -type l -exec ls -l 2>/dev/null {} +
lrwxrwxrwx 1 root root 24 Mar 18 12:41 /tmp/test/link -> /usr/admin/Collect_tests
lrwxrwxrwx 1 root root 43 Mar 18 12:41 /tmp/test/link1 -> /usr/admin/Collect_tests/Upload.CM@.www.com
lrwxrwxrwx 1 root root 68 Mar 18 12:41 /tmp/test/link2 -> /usr/admin/Collect_tests/Upload.CM@.www.com/Upload_Shema@@@.DATA.com
lrwxrwxrwx 1 root root 100 Mar 18 12:42 /tmp/test/link3 -> /usr/admin/Collect_tests/Upload.CM@.www.com/Upload_Shema@@@.DATA.com/List.files.emails.dummy*Printed
lrwxrwxrwx 1 root root 92 Mar 18 12:42 /tmp/test/link4 -> /usr/admin/Collect_tests/Upload.CM@.www.com/Upload_Shema@@@.DATA.com/List.files@emails.dummy
Examples of values
TARGET_NAME=Upload.CM@.www.com
TARGET_NAME=Upload_Shema@@@.DATA.com
TARGET_NAME=List.files.emails.dummy*Printed
Target: print: "link name" and "PATH" (last field ) only if $TARGET_NAME match exactly the word in the last field.
Example (when we want to match exact - while TARGET_NAME=Upload_Shema@@@.DATA.com then):
The results will display as the following
/tmp/test/link2 /usr/admin/Collect_tests/Upload.CM@.www.com/Upload_Shema@@@.DATA.com
/tmp/test/link3 /usr/admin/Collect_tests/Upload.CM@.www.com/Upload_Shema@@@.DATA.com/List.files.emails.dummy*Printed
/tmp/test/link4 /usr/admin/Collect_tests/Upload.CM@.www.com/Upload_Shema@@@.DATA.com/List.files@emails.dummy
There are a few conditions:
1) Need to match only the last field (from ls -l output)
Example
/usr/admin/Collect_tests/Upload.CM@.www.com
2) $TARGET_NAME value should match the whole word
Example of full match ( while TARGET_NAME=Upload.CM@.www.com ):
/usr/admin/Collect_tests/Upload.CM@.www.com
Example of a non-full match:
/usr/admin/Collect_tests/Upload.CM@.www.c
3) A backslash ("/") must exist on the left side of $TARGET_NAME, and a backslash or the end of the string must be found on the right of $TARGET_NAME.
4) Need to escape special characters as: " / " , " @ " . " * " , etc
5) The code will be part of a ksh script (and could beimplemented by a Perl oneliner or AWK or ksh etc .. )
Example
find /tmp -type l -exec ls -l 2>/dev/null {} + | < Perl one liner .............. >
As mentioned in response to your last question (since deleted), parsing ls output is very suboptimal. readlink can be used instead.
find /tmp -type l -exec \
perl -e'
my $TARGET_NAME = shift;
for (@ARGV) {
my $p = readlink($_);
$p =~ m{(?:^|/)\Q$TARGET_NAME\E(?:/|\z)}
or next;
print("$_\t$p\n");
}
' "$TARGET_NAME" {} \;
Or more efficiently,
perl -MFile::Find::Rule -e'
my ($TARGET_NAME, $BASE) = @ARGV;
for (File::Find::Rule->symlink->in($BASE)) {
my $p = readlink($_);
$p =~ m{(?:^|/)\Q$TARGET_NAME\E(?:/|\z)}
or next;
print("$_\t$p\n");
}
' "$TARGET_NAME" /tmp
As requested, this will match
TARGET_NAME
TARGET_NAME/
TARGET_NAME/x
.../TARGET_NAME
.../TARGET_NAME/
.../TARGET_NAME/x
but not
TARGET_NAMEx/...
.../TARGET_NAMEx
.../TARGET_NAMEx/...
xTARGET_NAME/...
.../xTARGET_NAME
.../xTARGET_NAME/...
Note: Change find ... -exec ... \; to find ... -exec ... + if your find supports it.