1

I need to replace many strings in many files using another file with patterns (like a database of strings). It is for anonymizing files.

Example:

File #1:

"Administrator";"512";"Built-in account for administering the computer/domain";"False";"False";"Administrator";"True";"True";"True";"S-1-5-21-3445027559-693823181-3401817782-500";"User";"OK";"23. 1. 2012 9:41:34";"20. 1. 2012 16:01:33";"10";"True";*

File #2 (the pattern file):

Guest;user1

Administrator;user2

system;user3

The output in File #1 needs to be

"user2";"512";"Built-in account for administering the computer/domain";"False";"False";"user2";"True";"True";"True";"S-1-5-21-3445027559-693823181-3401817782-500";"User";"OK";"23. 1. 2012 9:41:34";"20. 1. 2012 16:01:33";"10";"True";

Because the content of File #1 is variable, using some regex did not work in my case. I tried to use grep for finding, but it did not work either.

So far i tried this one

find file1.txt -type f -exec sed -f patt.regular {} \;

patt.regular is like this:

s/10.30.8.204/XX.YY.8.204/g

Unfortunately it does not work, I have suspicious, That source file format file1.txt have some issues, becouse nothing gonna happens, but when i delete original file (file1.txt) and create new one, it seems to be ok

Is there a possibility that sed could have some format trouble?

Andrea Corbellini
  • 17,339
  • 3
  • 53
  • 69
Lukas Toman
  • 33
  • 3
  • 2
    Please add what you tried so far, so we have something to start with. In addition, we may be able to point out what you got wrong in your own attempts. – Jongware May 11 '15 at 09:16

1 Answers1

1

In the shell, 

for line in $(cat file2.txt); do sed -i "s/$(echo $line | cut -d ';' -f 1)/$(echo $line | cut -d ';' -f 2)/g" file1.txt; done

does what you want. It needs the files to be called file1.txt and file2.txt and to reside in the current directory.

This does a sed for each line in file2.txt.

serv-inc
  • 35,772
  • 9
  • 166
  • 188