How to parse out fields using awk?

[Gino Castoldi_2]

Hi,

I have a shell script that finds and lists the full path to a file. (The filename always changes) I need to parse out all fields including the forward slashes because I only need the filename at the end of the path.
ex. /var/opt/filename
What would be the best way to do this?

10 points ot any good answer.
TIA, Gino

[harry d brown jr · Accepted solution]

man basename

live free or die
harry

[Jose Mosquera]

Hi,

Are you looking for the "basename" command?
#man basename

Rgds.

[Mark Grant]

This is confusing to me.

Firstly "basename /var/opt/filename" will give you "filename". However, I don't understand why you would need to do this because presumably you already had the filename in order to find it.

[Gino Castoldi_2]

Hi,

Mark: I'm using the 'find' command to locate the file so it returns the fullpath.

All: The 'basename' command looks like the solution to me.

10 points to any good answer.
TIA, Gino

[harry d brown jr]

Also dirname returns the directory:

# filevar=/var/adm/syslog/syslog.log
# basename $filevar
syslog.log
# dirname $filevar
/var/adm/syslog
#

live free or die
harry

[MarkSyder]

If you're using find, how about:

find / -name filename -exec basename {} \;

Mark Syder (like the drink but spelt different)

[Victor Fridyev]

Hi,

basename $FILE
and
dirname $FILE
are the best solution, but if you like awk:
PATH=$(echo $FILE|awk -F/ '{$NF="";print}' | sed 's? ?/?g'

HTH

[Fabio Ettore]

Hi Gino,

sorry, but the find command doens't already show the fullpath?
For example:

filename under /pippo/pluto
# find / -name filename
/pippo/pluto/filename

It returns fullpath then I don't understand your problem.
Otherwise I hope that other ITRC guys will help you better than me.

Best regards,
Ettore

[Gino Castoldi_2]

Hi,

Fabio: When I run the find command while searching for a particular file it returns the full path with the file itself.
I just need the file name by itself, I want to remove the rest of the string (which includes the full path).

10 points to any good answer.
TIA, Gino

[Fabio Ettore]

Ok, now I understand your problem.
I saw your first post but the second one diverted me:

***************
Mark: I'm using the 'find' command to locate the file so it returns the fullpath.
***************

No points for me....

Best regards,
Ettore

[Steve Post]

Ok.....
find /mydir -print | xargs basename?
NO. THAT won't work.
How about...
find /mydir | awk -F/ '// {print $NF}'
Yep. That looks right. I saw the awk command in the forums above.

But if you are looking for the base name perhaps you are just looking for the file?
If that's the case....."I want to find a file that begins with "billy".
find /mydir -name "billy*" -print | awk -F/ '// {print $NF}'

In awk the -F/ is means separate the fields by the / character.
The $NF means the current number of fields on that line.
So print $NF means print the last field on the line.

[Bill Hassell]

basename and dirname do a great job in parsing just the name and the path respectively. However, the shell already knows how to do this so you can bypass awk, sed and basename with something like this:

find /somewhere | while read
do
echo ${REPLY##*/}
done

[H.Merijn Brand (procura]

# perl -MFile::Find -MCwd=abs_path 'find (sub { my $filename = $_ ; my $relative_name = $File::Find::name; my $fullname = abs_path ($_) }, @ARGV)' dir dir dir

Modify to your liking

Enjoy, Have FUN! H.Merijn

[Tom Maloy]

If you are using ksh, you can use variable manipulation.

If the variable is stored in variable "fn", try this:

echo ${fn##*/}

and it should show you just the file name.

The "##" removes the large left pattern.
The "*/" is the pattern to be removed.

Tom

[Muthukumar_5]

Hai.

You can get the directory name and pathname of all regular files using,

#!/usr/bin/ksh

echo " DIR-NAME FILE-NAME"
find / -type f | while read file; do
echo "`dirname $file` `basename $file`"
done

For other files like character files,block devices change the -type X value from man find

Regards,
Muthukumar.