Shell Programming and Scripting

Hi All,

I am trying to manipulate a file using awk in UNIX. My file "test_file" is structured as follows:

field1:field2:field3
field4:field5:field6
field7:field8:field9

My requirement is to calculate another field (which would be let's say sum of 2nd and third column and put as fourth column in this file). so that my final output would be

field1:field2:field3:fieldA
field4:field5:field6:fieldB
field7:field8:field9:fieldC

where fieldA=field2 + field3 and so on...

I have the following piece of code which I guess, is nowhere in the working condition right now:

while read line
do
echo $line | awk -F: 'BEGIN {OFS=":"}
{$4=$2 + $3 ; print $0}' >> new_test_file
done < test_file

I read somewhere that you can dynamically calculate variables in awk, so $4 would be dynamically calculated at the run time. Subsequently when i print $0 should contain the new calculated field too. But I guess that is not the case.

Can anybody look into above and help me out?

Thanks

Guyz,
Plz ignore my post. I just tried it on a different unix box and it worked. Both are korn shells. I have no idea why it did not work on the first one. I am going to do some further R&D on this.
Thanks anyways.
Vikas.

Guys

I finally resolved the problem. As i mentioned in my last post that somehow, this code was working on one box and giving prob on second one. Well, i try the same code with "nawk" on second one and it worked too.

Now, could someone plz explain why this happened? Can I continue to safely use nawk on my second server or is it not advisable. Appreciate any response, ppl.

Thanks
Vikas.

I don't know anything about nawk, but your use of $4 seems wierd to me. $4 should be the 4th field in the record, but in your example there is no 4th field. Instead, you want to create the 4th field. Also, why are you bothering with a read?

I would skip the read and do something like --

awk -f myawk.awk myfile.txt > mynewfile.txt

Code:

# myawk.awk
BEGIN {
 FS=":"
 OFS=":"
}

{ 
 fld4 = $2 + $3
 print $0, fld4
}

END {}

Hi Glenn
Thanks for your reply. However, $4 can definitly be calculated at the runtime. As I said i ran this piece of code on some other unix box and it ran fine. Also if i use nawk, it runs fine on the first one too.

Anyways, as I said, awk and nawk can dynamically calculate variables at run time. So even though i have $4 undefined, but i can calculate it at run time.
Secondly, i have to use read in a while loop, since apart from calculating a field like above, i am doing some other operations too. Lastly, $4 was just an example. My file could be a variable length file too in which case, i would need to use $NF to calculate the last field and then probably i can use $(NF+1) to calculate the new field and embed in the new file.

As far as nawk is concerned, i guess it stands for new awk. I read in some forums, that there is a difference in version of awk. So my best guess is my first box has an older version of awk whereas second one has a newer version. That is why i have to new awk or nawk on my first machine.

Thanks for your reply again
Vikas.