[SOLVED] using for loop in awk

ilukacevic · 03-27-2011, 01:59 PM

Dear all,

my situation is following: I would like to use the for loop (not necessarily for loop) with awk command in such a way that awk doesn't evaluate first the first row through all the iterations of the loop, but to evaluate all the rows for the first iteration, then all the rows for the second iteration, and so on.

For example, if I have a file like

2 3
4 5

and I want to print $1,$2,$1*t,$2*t, for the values of t from 1 to 3, I would like to get

2 3 2 3
4 5 4 5
2 3 4 6
4 5 8 10
2 3 6 9
4 5 12 15

and not

2 3 2 3
2 3 4 6
2 3 6 9
4 5 4 5
4 5 8 10
4 5 12 15

what I always get with my tries. Two of my tries are

awk '{ for (t=1; t<=2; t++) print $1,$2,$1*t,$2*t}' proba.dat > proba1.dat

awk '{ t=1
while (t<=5) {
print $1,$2,$1*t,$2*t
t++
}
}' proba.dat > proba2.dat

Does anyone have a solution to my problem? I would appreciate any hints and help.

Thank you in advance!

Yours,
Igor

Nominal Animal · 03-27-2011, 03:32 PM

You'll need to read the input into an array, and do the output in a loop at end:

Code:

awk '{ line[NR] = $0 }
 END { for (t = 1; t <= 3; t++)
           for (i = 1; i <= NR; i++) {
               printf("%s", line[i])
               n = split(line[i], field)
               for (k = 1; k <= n; k++)
                   printf(" %g", t * field[k])
               printf("\n")
           }
     }' proba.dat > proba1.dat

Does this solve your problem?

colucix · 03-27-2011, 04:04 PM

There are different ways to perform a loop using awk. You can rewind the file to parse it multiple times, or you can pass the file name as argument multiple times. Some options are covered in this thread, see in particular post #2 and post #6. The suggestion by Nominal Animal is good, unless the file size is huge.

kurumi · 03-27-2011, 07:52 PM

I would recommend using a tool that can seek and tell a file (Python|Perl|Ruby etc) . an example in Ruby

Code:

#!/usr/bin/env ruby 
# Ruby1.9+
f=File.open('file')
1.upto(3) do |i|
    while not f.eof
        s = f.readline.split.map(&:to_i)
        puts "#{s[0]} #{s[1]} #{s[0]*i} #{s[1]*i}"
    end
    f.pos = 0 # goes back to start of file...
end
f.close

example run

Code:

$ ruby myscript.rb
2 3 2 3
4 5 4 5
2 3 4 6
4 5 8 10
2 3 6 9
4 5 12 15

ilukacevic · 03-28-2011, 01:09 AM

Dear Nominal Animal and kurumi,

both of your solution work well. Thank you both!

The only issue I have is with the ruby script...I don't know how to pipe the output to a file. I'll try to find it on the web.

Yours,

Igor

kurumi · 03-28-2011, 02:15 AM

Quote:

Originally Posted by ilukacevic

The only issue I have is with the ruby script...I don't know how to pipe the output to a file. I'll try to find it on the web.

Yours,

Igor

you can pipe the file when you run the script because there is a command called puts (print to screen)

Code:

$ ruby myscript.sh > new file

Or you can do it inside the script itself by opening output file handle. ( You can check it out yourself on how to do it)

ilukacevic · 03-29-2011, 02:48 PM

Quote:

Originally Posted by kurumi

you can pipe the file when you run the script because there is a command called puts (print to screen)

Code:

$ ruby myscript.sh > new file

Or you can do it inside the script itself by opening output file handle. ( You can check it out yourself on how to do it)

OK. Thanks, again!

Igor