renaming (or appending to) only a part of a string

[Osman Yagmurdereli]

Hello again.

I need some ideas about renaming some part of a string.

I have text file which have directory names in it, like this:

/usr/bin/dummyfile.dat
/var/local/src/file.finder.2000.dat
/var/spool/printerqueue.dat
/home/ICS/U990702043/The man who saves the world.dat
...
...
...

I need to append a different string to these strings and save the results to a different or the same text file. Appending (or renaming) must be on the last name, or after the last slash (/). For example, I want to append 'here_is_' string. Results should be like:

/usr/bin/here_is_dummyfile.dat
/var/local/src/here_is_file.finder.2000.dat
/var/spool/here_is_printerqueue.dat
/home/ICS/U990702043/here_is_The man who saves the world.dat

If someone would help I would be most grateful.

[Osman Yagmurdereli]

should I use awk, sed, or cut? how can I append after the last '/'? I am so confused...

[Dennis Handly]

It seems you need to use awk to see where to
insert that string. Your above paths seem to have differing number of "/" and you want the last one.

>how can I append after the last '/'?

Just use substr in a loop to find the last one.

[Denver Osborn · Accepted solution]

There are several ways it could be done, this is just one.

while read line
do
print `dirname "${line}`/here_is_`basename "${line}"`
done < infile.txt


Hope this helps
-denver

[Denver Osborn]

sorry, forgot a "

:)


while read line
do
print `dirname "${line}"`/here_is_`basename "${line}"`
done < infile.txt

[Peter Godron]

Osman,

#!/usr/bin/sh
string=`echo "here_is_" |rev`
rev dat.lis | sed "s/\//${string}\//" | rev

Method:
Reverse the data to be inserted
For each line in the file dat.lis
reverse the line
subsitute the (now first) instance of '/' with the new data
reverse the line to 'normal'


Please also read:
http://forums1.itrc.hp.com/service/forums/helptips.do?#33 on how to reward any useful answers given to your questions.

[Peter Godron]

Osman,
just for the choice, here is another way:
sed 's/\(.*\)\//\1\/here_is_/' dat.lis

[Sandman!]

Try the awk construct below:

echo | awk -F/ '{
for(i=1;i p=(p?p"/"$i:$i)
str=p"/here_is_"$NF
system("mv "$0" "str)
}'

[Peter Nikitka]

Hi,

you get a more readable version of Peter G.'s solution in NOT using '/' as sed command-delimiter:

sed 's:\(.*\/):\1here_is_:' dat.lis

mfG Peter

[Matti_Kurkela]

Here's a shell solution:

#!/bin/sh
while read oldname
do
directory=$(dirname "$oldname")
filename=$(basename "$oldname")
newname="${directory}/here_is${filename}"
echo "$newname"
done

Pipe the text file into this script, and you'll get the modified version to standard output.

For example:
sh script.sh modified.txt

[Matti_Kurkela]

Oh, sorry, seems Denver Osborn already showed this way.

No points for me from my previous answer, thanks!

[Peter Godron]

Hi Osman,
you have a whole range of solutions here. One-liners to mini-scripts.

If these answers helped you with solution(s) could you please complete the thread by awarding points to helpful answers and summarising the solution for you.

This will help resolution of similar problems in the future.

[Osman Yagmurdereli]

of course, i will.

sorry for the delay.