I wouldn't use grep. As the other answers already said, it's not the right tool for this job.

Considering your specific case (fields separated by /), basename is the most straighforward way, as stated in Dirk's answer. I just want to provide some other alternatives.

cut

The cut command can separate the input in fields and allows you to get only the ones you need. If you know for sure that the inputs will always be branches/features/something, then you can use / as separator and get the third field:

echo $CCSR | cut -d / -f 3

Or, if you don't know how many fields there will be, there's a little trick to get the last one:

echo $CCSR | rev | cut -d / -f 1 | rev

rev reverses the input, so first I reverse it, get the first field and reverse it again. The result will be the last field.

awk

The awk command works in a similar way: we tell it what the separator is and can get specific fields. The difference is that it has a better way to get the last one, by using the predefined variable NF:

echo $CCSR | awk -F / '{print $NF}'

Although basename is the most straighforward solution for your case, please note thatcut and awk are more generic, as they can work with different separators and allows you to get specific fields (not only the last one).

posted over 2 years ago

CC BY-SA 4.0

2y ago

hkotsubo‭

5020 reputation 20 68 566 221

grep is not the right tool for your case. You can use basename:

basename a/b/c   -->   c
basename a/b/c/  -->   c

or, in your case

basename branches/features/arm_and_musl  -->  arm_and_musl

which, within a sh script, you would use as

dirname=`basename "$CCSR"`

Note, that if you need to call sh explicitly, you have to give the -c command line argument to sh if you want to pass commands on the command line.

posted over 2 years ago

CC BY-SA 4.0

Dirk Herrmann‭

1387 reputation 1 31 141 48

While I agree with Dirk Herrmann that basename is the right tool for the job, I think it is still worthwhile to know how to do it with grep, because you might one day encounter a sufficiently similar situation that is not covered by a specialised tool.

The question can be broken into two questions:

1. What regular expression matches exactly the part after the last slash.

2. How to get grep to only output the part that matches, instead of the whole line.

For the first question, note that the part after the last slash is characterized by the fact that it doesn't contain a slash (otherwise the previous slash would not have been the last one) and extends to the end of the line.

To match a character that's anything but a slash, you can use [^/]. You want to match any number of them, so put a star afterwards: [^/]*. But you only want to match it if it extends to the end of line (so you do not match e.g. the branches in your example), this is done by terminating the regular expression with a dollar sign. So the complete regular expression you need is [^/]*$

For the second question, the option -o does exactly what you need. From the man page:

       -o, --only-matching
              Print  only  the  matched  (non-empty) parts of a matching line,
              with each such part on a separate output line.

So putting it all together, we get the desired behaviour:

$ echo "branches/features/arm_and_musl" | grep -o '[^/]*$'
arm_and_musl

posted over 2 years ago

CC BY-SA 4.0

celtschk‭

771 reputation 8 12 97 2