awk parsing

[Thi Vu]

Hi all,

I got a list of file like this:

mcdbq/ABCDGEF.P
crtbn/GIFGEO.I
djgei/DKJFEO.VAR
rtbkdjf/fjiefj.p
rgmien/fjeijg.i
rteji/fjeij.d

The .P can be whatever as long as it's in capital. I am trying to get a listing of the files in capital only using this command:

tar -tvf filename.tar |awk -v file="[A-Z]*\.[A-Z]*$" '{if($8 ~ file)print $8'

It prints all the files (including the non-capital file); but if I do this:

tar -tvf filename.tar |awk -v file="[A-Z]*\.P*$" '{if($8 ~ file)print $8'

it only prints the .P file, which is not all the file I want

I even use this command:

tar -tvf filename.tar |awk -FS="/" -v file="[A-Z]*\.[A-Z]*$" '{if($8 ~ file)print $8'

and there is no output at all. Where did I got wrong? Any help is appreciate.

Thi

[RAC_1]

Not AWK, but simple grep.

tar -tvf filename.tar | grep -e "[A-Z]\$"

Anil

[Hein van den Heuvel · Accepted solution]

Here is one way:

awk '{n=split($0,p,"/");file=p[n]; if (file == toupper(file)) { print file}}' your-list-file

split line into parts based on "/"
filename it the last part ("n")
print if filename equals uppercased filename.

Hein.

[curt larson_1]

give this a try

tar -tvf filename.tar |awk -v file="\.[A-Z]+$" '{if($8 ~ file)print $8'

use a + instead of *
* matches zero or more, as in it doesn't have to be there
+ matches one or more, there has to be at least one.

[Thi Vu]

Hi Anil and Curt,

Your commands show everything (small letter files and capital letter files)

Hein,

Your command work, however thinking further down at the script I would like to know how do I go about if I want this:

I have: mcstm/ABDEGD.P

I would like to assign file=ADBEGD.P and
dir=mcstm where I will call the variable file and dir several times somewhere else in the script. Thanks

Thi

[RAC_1]

tar -tvf filename.tar | grep -e "[A-Z]\$"|awk -F "/" '{print $1, $2}'

Anil

[Rodney Hills]

The reason it is not doing what you want is the way "\" is being intrepetted by the shell and awk.

Make the following change (double quotes to single and an extra "\")-

tar -tvf filename.tar |awk -v file='[A-Z]*\\.[A-Z]*$' '{if($8 ~ file)print $8'

HTH

-- Rod Hills

tar -tvf filename.tar |awk -v file="[A-Z]*\.[A-Z]*$" '{if($8 ~ file)print $8'

[Michael Schulte zur Sur]

Hi,

tar -tvf filename.tar | awk -v file="\.[A-Z]+$" '{if($8 ~ file)print $8}' | while read FILE; do set file=`basename $FILE`;set dir= `dirname $FILE`;done;

Is that what you were looking for?

greetings,

Michael

[Hein van den Heuvel]

> Your command work, however thinking further down at the script

In the awk script or in a shell script that is using n awk command? Check out Micheals example of how to get awk output in a symbol.
In my example the symbols ar ether within the awk script for further processing.
Here are some more samples:

awk:

tar -tvf | awk '{if (match($8,/\/[A-Z\.]+$/)){ dir=substr($8,0,RSTART-1); file=substr($8,RSTART+1,999); print dir,file}}'

perl:

tar -tvf | 'if (/\s(.*)\/([A-Z\.]+)$/) { print "dir=$1, file=$2\n"}'

Hein.