how to shorten paths in files

GermanicGalore

Hi everyone,

I am new to shell scripting and I am trying around forever now with sed and all sorts of stuff to do something that sounds really simple.

I have a directory with a couple of files. The files contain a full url path and I need to shorten that.

Example:

"https://my.full.path/myFile"

needs to be shortened to

"myFile"

The files contain more than just urls in them. So, I need to search for the pattern "https://my.full.path/myFile", then replace it in the same file with "myFile". There are a couple of hundred files in that directory.

Sounds simple enough, but I cannot seem to get it together.
Sorry for the newbie question!

croo

It's while since I was writing any serious scripts, but when I was this type of problem was resolved using Pattern-Matching Operators... that was with the korn shell but I think bash is compliant with it.

Just tested and yes it works too in bash.

To get myFile from "https://my.full.path/myFile"
You would do something like

#!/bin/bash
path="http://my.full.path/myFile"
echo ${path##http://*/}

Google I am sure can find you the details of the operators. I think there were just 4 variants.

E39MSport

in ksh I'd probably use

echo "https://my.full.path/myFile" | sed 's/^.*\/\(.*\)$/\1/'

ok unless you have millions of records. perl would possibly be handier for the latter

croo

ok I did a quick google and did not see any links to what I meant on the first screen so here is very short & quick overview.

There were 4 "Pattern-Matching Operators" that I remember

${variable#pattern} - If the pattern matches the start of the variable's value, delete the shortest part that matches and return the rest.
e.g.

path=/home/croo/test/long.file.name
${path#/*/} --> croo/test/long.file.name  

${variable##pattern} - If the pattern matches the start of the variable's value, delete the longest part that matches and return the rest.
e.g.

path=/home/croo/test/long.file.name
${path##/*/} --> long.file.name

${variable%pattern} - If the pattern matches the end of the variable's value, delete the shortest part that matches and return the rest.
e.g.

path=/home/croo/test/long.file.name
${path%.*} --> /home/croo/test/long.file

${variable%%pattern} - If the pattern matches the end of the variable's value, delete the longest part that matches and return the rest.
e.g.

path=/home/croo/test/long.file.name
${path%%.*} --> /home/croo/test/long

And that's it.. they are short & simple but very powerful!

pickarooney

Can't you just use the basename command or am I missing something?

stesh

Try the cut command:

% echo "https://my.full.path/myFile" | cut -d '/' -f 4
myFile

Here -d specifies the delimiter, and -f specifies which portion of the string you want to retain. You can also select multiple portions:

 echo "https://my.full.path/myFile" | cut -d '/' -f 2,4
/myFile

(portions 2 and 4)

% echo "https://my.full.path/myFile" | cut -d '/' -f 2-4
/my.full.path/myFile

(the portions between 2 and 4, inclusive)

E39MSport

But then OP is constrained to a fixed path length/format?

jonathan11

Nice feature in ksh for doing this...

##### Strip everything before last slash:
path="https://my.full.path/myFile"
echo ${path##/*/}

(as croo said above)