[SOLVED] bash - Copying two columns and doing an operation to the third

calebante · 05-25-2013, 01:41 PM

I have an executable "a.out" that does an operation with a number, for example duplicates it.

$./a.out 12
24

and I have a test file "file.dat" like this:

11 12 13
21 22 23
31 32 33
41 42 43

I'd like to obtain a file identical to the original, except by the 3rd column, in which every value has been replaced by its double, found by a.out:

11 12 26
21 22 46
31 32 66
41 42 86

I'm new in shell scripting and I'm totally confused. I would appreciate a lot any help!

grail · 05-26-2013, 12:09 AM

Well as you have mentioned scripting I will go on a limb (as you don't say) and go with bash. To this end you will need the following:

1. A loop to read in the original values

2. Call your a.out on the last value

3. Redirect the output to a new file

If you do not have a reference, try http://tldp.org/LDP/abs/html/

Let us know how you get on

David the H. · 05-26-2013, 07:46 AM

This kind of operation is trivial to do in awk, by the way, without even the need for any external program. I just accomplished it myself with an expression that's a whole 8 characters long.

Here are a few useful awk references:
http://www.grymoire.com/Unix/Awk.html
http://www.gnu.org/software/gawk/man...ode/index.html
http://www.pement.org/awk/awk1line.txt
http://www.catonmat.net/series/awk-one-liners-explained

grail · 05-26-2013, 09:18 AM

hmmm ... I can go 5 if not using a.out and assuming we do not include the quotes

danielbmartin · 05-26-2013, 10:46 AM

Quote:

Originally Posted by grail

I can go 5 if not using a.out and assuming we do not include the quotes

Please post this code.

Daniel B. Martin

grail · 05-26-2013, 10:50 AM

Code:

awk '$3*=2' file

ntubski · 05-26-2013, 10:57 AM

What if you have a 0 in column 3? I think 8 characters is the minimum for a correct solution.

calebante · 05-26-2013, 11:08 AM

I know that

awk '{print $1,$2,2*$3}' test.dat

would work if I just wanted to double the 3rd column, but this is not what my a.out does. I just gave this duplication thing as an example...

For each line, I have to write the 1st value, write the 2nd value, and write the value returned by a.out on the 3rd value!

druuna · 05-26-2013, 11:25 AM

Have a look at this:

Code:

awk '{ printf("%s %s ",$1,$2) ; system("./a.out "$3) }' test.dat

If a.out is not in the directory you execute this command from, prepend the full path to it.

grail · 05-26-2013, 12:10 PM

I concede to the zero

calebante · 05-26-2013, 01:24 PM

Yes, it works!

Great, druuna! Thank you all of you.

David the H. · 05-28-2013, 10:30 AM

To do it in the shell, BTW, you just have to split the data properly as you read it.

Code:

while read -r -a fields ; do

    fields[2]=$( mycommand "${fields[2]}" )
    echo "${fields[*]}"

done <infile.txt

This example assumes you're using a shell like bash with array support, but it could easily be converted into posix-compliant code as well.