Hi There,

Assume there is file call version.txt which contains following text line.

090df276848c2fb2fbdd0e7449284768ead8348b commit refs/tags/PNB.X.X.3.015.0

I wont to grep this version.txt file and get the result only the part of the version "PNB.X.X.3.015.0"

Note, this "PNB.X.X.3.015.0" is dynamically changing and the "PNB" part of the version is still renaming same.

Please help.

grep "090df276848c2fb2fbdd0e7449284768ead8348b commit refs/tags/PNB" <filename> | awk -F/ '{print $NF}'

You encapsulate the string in double quotes since white space is the default field separator. You end at PNB since you say what follows that changes.

You pipe into awk and use the -F flag to specify field separator to be "/" rather than white space. The NF is a special variable in awk meaning "number of fields" so $NF is reference to the (last) number. Since your PNB... is preceded by "/" it is a perfect separator to use for what you asked.

1 members found this post helpful.

You could use sed for this:

This command will get the single line from the file and removes anything until "tags/" (including) and returns the remaining line (your version).
If you add the parameter '-i' sed will change the file in place.

1 members found this post helpful.

Thank You so much.
Its giving me the answer what I need. could you please further clarify me about the syntax which you have use here.

awk -F/ '{print $NF}

Hope this would be a good learning point for me. appreciate it.

Thank you so much. both answers has given what I need.

I edited my original post to explain it. Let me know if you need more.

1 members found this post helpful.

Grep has "-o" to return only the requested text. See the manpage.

Thank You all,

how can I write the output of the command into a file as a variable.

for example, how can I wire the following value into a file named "app.version.txt"


version=PNB.X.X.3.015.0

So the "version=" part should be added before the "PNB.X.X.3.015.0"

This assumes he has no other lines with *tags* on them but is a nice solution.

1 members found this post helpful.

Of course since he says what follows PNB changes he'd have to do something like:

grep -o PNB.......... <filename>

But like the sed solution that assumes no other lines have PNB in them and also assumes that the number of characters following PNB are always the same.

1 members found this post helpful.

Can you help me on this? people like you help us to save our time.
I would like to learn all this, but I just confuse about the command line "sed"


Thank You all,

how can I write the output of the command into a file as a variable.

for example, how can I wire the following value into a file named "app.version.txt"


version=PNB.X.X.3.015.0

So the "version=" part should be added before the "PNB.X.X.3.015.0"

That's what regex was invented for.

No more fragile than your presumption there are no other lines with both a slash and PNB in a different order.

Sorry for asking silly question. I have done this using below.
#!/bin/bash

git_tag=`git for-each-ref --sort='-committerdate' 'refs/tags' | head -n 1 | awk -F/ '{print $NF}'`
echo "version="$git_tag > version.txt

When using bash I'd suggest instead:
The `` certainly works but is the old style. Using $() has a couple of advantages:
1) ` is often confused with ' when reading lines so debugging is sometimes a problem.
2) The $() can be nested more easily than the ``.

1 members found this post helpful.

I didn't assume - I took the OP at his word. He said the only thing that changes is what is after PNB.

There was no slur intended against your suggestion - I was just trying to spell it out further for the OP.

1 members found this post helpful.