CODEWITHSUNDEEP
linuxunixsedawk
Mark McWhirter
Mark McWhirter

Reputation: 1196

How to extract filename from path using sed or awk

I am trying to parse a filename from a modified apache web access log entry that is tab delimited:

/common/common/img/pictos/klArrowRight.gif    /common/common/img/pictos/klArrowRight.gif   03/Dec/2012:00:00:00    127.0.0.1   03/Dec/2012:00:00:00    us   404

I would like it to come out like this:

klArrowRight.gif    /common/common/img/pictos/klArrowRight.gif   03/Dec/2012:00:00:00    127.0.0.1   03/Dec/2012:00:00:00    us   404

I have tried something like this in :

's:.*/::'

However, it is too greedy, and it eats the rest of my line. I have been looking through posts, but so far no luck. Any hints?

Upvotes: 4

Views: 22118

Answers (6)

user12531
user12531

Reputation: 81

None of the given answers seem to be correct completely when only the extraction of a filename from a given absolute path is desired. Therefore I give here the solution. Let's consider in variable filename we have the complete path, e.g., filename=/ABC/DEF/GHI Then,

echo $filename | awk 'BEGIN{FS="/"}{print $NF}'

will result in the filename GHI.

Upvotes: 8

Steve
Steve

Reputation: 54392

One way using GNU grep:

grep -oP "[^/]*\t.*" file

Results:

klArrowRight.gif    /common/common/img/pictos/klArrowRight.gif  03/Dec/2012:00:00:00    127.0.0.1   03/Dec/2012:00:00:00    us  404

Upvotes: 0

Ed Morton
Ed Morton

Reputation: 203368

awk 'BEGIN{FS=OFS="\t"} {sub(/.*\//,"",$1)} 1' file

Upvotes: 1

Norman Gray
Norman Gray

Reputation: 12514

You can do this pretty easily with just sed, as long as you tell it not to be too greedy:

% echo '/img/pictos/klArrowRight.gif 03/Dec/2012' | sed 's,^[^ ]*/,,'
klArrowRight.gif 03/Dec/2012
%

(that is, "starting at the beginning of the line, find the longest-possible list of non-space characters, followed by a slash")

Upvotes: 4

zb'
zb'

Reputation: 8059

using perl regexp and basename (i not think you stuck on sed/awk): 

perl -p -e 'use File::Basename;s/([^\s]+\s+)[^\s]+\s+/$1/;print basename($1)'

example: 

echo "/common/common/img/pictos/klArrowRight.gif /common/common/img/pictos/klArrowRight.gif 03/Dec/2012:00:00:00 127.0.0.1 03/Dec/2012:00:00:00 us 404" |
   perl -p -e 'use File::Basename;s/([^\s]+\s+)[^\s]+\s+/$1/;print basename($1)'

klArrowRight.gif /common/common/img/pictos/klArrowRight.gif 03/Dec/2012:00:00:00 127.0.0.1 03/Dec/2012:00:00:00 us 404

Upvotes: 2

Kent
Kent

Reputation: 195049

the input/output in your question is not well formatted. do you need this?

awk '{gsub(/\/.*\//,"",$1); print}' file

test

kent$  echo "/common/common/img/pictos/klArrowRight.gif /common/common/img/pictos/klArrowRight.gif 03/Dec/2012:00:00:00 127.0.0.1 03/Dec/2012:00:00:00 us 404"|awk '{gsub(/\/.*\//,"",$1); print}'

output:

klArrowRight.gif /common/common/img/pictos/klArrowRight.gif 03/Dec/2012:00:00:00 127.0.0.1 03/Dec/2012:00:00:00 us 404

Upvotes: 2

Related Questions