The BASH scripting language provides a very handy ‘loop’ features. we have discussed that we can chain certain commands such as:
strace -p `pgrep -u root sshd`
This command however, provided that pgrep returns more than 1 process, will result in an error, since strace does not expect more than 1 process at a time. To overcome this issue, we can edit the line and add a loop control structure:
for i in `pgrep -u root sshd`; do strace -p $i ; done
What the above line will do is to create a for loop and for each result of pgrep to run strace -p.
Another example would be to use the same for statement to remove a certain line in each file that contains it. Let’s imagine that you have hundreds of files containing a footer with copyright message with wrong year (i.e 1998 for example) and you need to remove it. The problem is the files are numerous and it will take a lot of time to do it by hand. Fortunately, you can use a single line in BASH to do it for you.
For the purpose, we will create a for loop, which will grep the whole directory and its subdirectories for the message we need to remove and pass it to a SED command which will actually process the file and remove it. The full command line will look like this:
for i in `grep * -r -l Copyrighted`; do sed '/Copyrighted/d' -i $i ; done
The line broken in parts is as follows:
1)for loop: for i in something; do something; done;
2)grep * -r -l Copyrighted -- greps the message and returns only the file names containing it
3)sed ‘/Copyrighted/d’ -i $i -- uses sed to remove the line from the files we have located with the above command.
