getting the last field of a string separated by a particular delimiter in bash

Question

I'm writing a script that copies a set of files to a directory using bash scripting. A file that I'm trying to copy might already exist in the directory so I have to rename the file to avoid replacing the original one.

For an example if I have to copy a file called "abc.dd.java" and a file with the same name already exists, I have to rename the file as "abc.dd1.java" and copy it.

If the destination contains "abc.dd.java" and "abc.dd1.java" then the file should be renamed "abc.dd2.java"

I wrote a code snippet that does that:

#$file contains the full path for the original file.
#$path contains the full path for the copied file.

copypath="$path"
echo "$path"
count=1
while [ -f "$path" ]
do
    part1=`echo "$copypath" | awk 'BEGIN{FS="."}{for(i=1;i<NF-1;i++){printf("%s.",$i)}printf("%s",$(NF-1))}END{}'`
    part2=`echo "$copypath" | awk 'BEGIN{FS="."}{printf("%s",$NF)}END{}'`
    path="$part1""$count.""$part2"
    count=`expr "$count" + 1`
done
cp "$file" "$path"

This does work correctly. But it seems too crude. Isn't there a better way of separating out the file extension part and the name parts?

Thanks in advance

Possible duplicate: http://stackoverflow.com/questions/965053/extract-filename-and-extension-in-bash . But watch out for files without an extension! — tom, Oct 28 '14 at 08:55

score 2 · Answer 1 · answered Oct 28 '14 at 09:17

2

To Split the filename and extention you can use the below string manipulations

filename="/tmp/example.txt"

# will output "txt" (extention)
echo ${filename##*.} 

# will output "/tmp/example" 
echo ${filename%%.*}

answered Oct 28 '14 at 09:17

Ram

1,115
8
20

Test it with `"/tmp/another.example.txt"`. The second expansion should have `%` instead of `%%` (shortest match instead of longest match). – tom Oct 28 '14 at 09:21

score 1 · Accepted Answer · edited May 23 '17 at 12:27

A similar question has been answered in Extract filename and extension in bash.

Here is a simple version which (mostly) works, based on that link:

part1="${copypath%.*}"            # remove file extension (shortest .* suffix)
part2="${copypath##*.}"           # remove base path (longest *. prefix)
path="${part1}${count}.${part2}"

And here is a more complete answer which also handles files without extensions:

if [[ "$copypath" =~ [^/.][.][^/.]+$ ]]
then
    # copypath has an extension
    part1="${copypath%.*}"
    part2=".${copypath##*.}"
else
    # no valid extension, so insert the number at the end
    part1="$copypath"
    part2=""
fi
path="${part1}${count}${part2}"

Features:

Whitespace is handled well.
Multiple dots are handled as requested.
e.g. abc.dd.java becomes abc.dd1.java
Files without an extension are handled properly.
e.g. file becomes file1
Dot files with no extension are also handled properly.
e.g. .bashrc becomes .bashrc1

Caveats:

Multiple extensions are not handled specially.
e.g. archive.tar.gz becomes archive.tar1.gz instead of archive1.tar.gz
If the part before the extension ends with a dot it is not considered to be an extension.
e.g. This is a sentence..txt becomes This is a sentence..txt2
To treat this as having an extension change the regular expression to [^/][.][^/.]+$
Does not attempt to handle directories.
Only tested in Bash, unlikely to work with other shells (but could be changed to use grep).

score 0 · Answer 3 · answered Oct 28 '14 at 08:57

0

Try if you have just .java files with something:

fname=$(basename file.java .java)
echo $fname

output:

file

and then you can do your further processing.

answered Oct 28 '14 at 08:57

SMA

36,381
8
49
73

getting the last field of a string separated by a particular delimiter in bash

3 Answers3