mjhjmtu

I would like to copy line in text file edit it on the fly to replace all / to | and paste in on the same line in the same line.

I have the first line in the text file:

http://webpage.com/(name-never-changes)/aCcboeasdfdRD/asdfasdft21.jpg

Copy text between (name-never-changes) and .jpg

/aCcboeasdfdRD/asdfasdft21

Change to / to |

|aCcboeasdfdRD|asdfasdft21

And add to the end of the same line where URL is, with a separator
the separator can be text or number or special character

http://webpage.com/(name-never-changes)/aCcboeasdfdRD/asdfasdft21.jpg (separator) |aCcboeasdfdRD|asdfasdft21

And this should work with multiple text files with different URLs

I can't create new files and merge them this should be on the fly because the name of text files are different and I don't know filenames

Try this,

awk -F '[/.]' '"$(NF-2)"' file

http://webpage.com/(name-never-changes)/aCcboeasdfdRD/asdfasdft21.jpg (saperatot) |aCcboeasdfdRD|asdfasdft21

• 
  /. two field separators


• 
  $0 to print the complete line


• 
  $(NF-2) $(NF-1) the third and second fields from the end

if you want to save in the file

Option 1: (if it has few lines)

echo "`awk -F '[/.]' '"$(NF-2)"' file`" > file

Option 2: (if it is a large file)

 awk -F '[/.]' '"$(NF-2)"' file > tempFile ; mv tempFile file

For multiple files in a folder:

cd /path/to/dir
for file in `ls`
do
 echo "`awk -F '[/.]' '"$(NF-2)"' $file`" > $file
done

Here’s a sed solution:

sed 'h; s@^[^/]*//[^/]*/[^/]*/@ (separator) |@; s@/@|@g; s/.jpg$//; x; G; s/n//'

Step by step:

• 
  h copies the current line from the “pattern space” to the “hold space”.


• 
  s@^[^/]*//[^/]*/[^/]*/@  (separator)  |@ is a substitute command
  using @ as the delimiter. 
  It replaces

  stuff//stuff/stuff/

  
  (i.e., http://example.com/(top-level-directory)/)
  with   (separator)  |.
  
  You could do
  s@^http//example.com/(top-level-directory)/@ (separator) |@
  if you wanted to.


• 
  s@/@|@g is also a substitute command using @ as the delimiter. 
  It replaces all remaining / characters in the line with | characters. 
  If you’re sure that there will be only one
  (between (second-level-directory) and (filename)),
  you can leave off the g.


• 
  s/.jpg$// obviously removes the .jpg extension from the end of the line. 
  (If you don’t know what the extension is,
  you could do something like s/.[^.]*$// instead.)


• 
  x exchanges the pattern space and the hold space,
  so the original line is now back in the pattern space, and the

    (separator) |(second-level-directory)|(filename)

  that we just built
  is in the hold space.


• 
  G gets the contents of the hold space
  and appends it to the pattern space (injecting a newline between them).


• 
  s/n// removes that newline.

This will pass blank lines through as blank lines. 
Other lines will be mangled.

If you want to edit file(s) in place, just pass sed the -i option
(or -i.bak, if desired or required on your system).