[SOLVED] awk variable

jkobori · 10-21-2012, 06:20 AM

Hi everybody,

I have the following code:

#!/bin/bash
.
.
.
for ((i=0;i<$detval-1;i++))
do
t="${array[i+1]}"
cat "${array[k]}"n"${array[i+1]}".d64 | awk '{print $t}' > fg.dat
done

The script supposed to cd to a particular directory, cat a file, then choose one column (t), and print it to fg.dat.
The problem is, that if I use "$t" in the awk, it does not do what I want it to do,
I guess beacuse I actually should use "$($t)". Of course, I've tried it, also cut -f $t,
but none of them has worked.
Can anyone help me out?
Thank you very much!

Regards,
Joe

David the H. · 10-21-2012, 07:35 AM

Please use ***[code][/code]*** tags around your code and data, to preserve the original formatting and to improve readability. Do not use quote tags, bolding, colors, "start/end" lines, or other creative techniques.

awk has a separate variable system from the shell, and shell values need to be passed into awk variables before they can be used internally. This is usually done with the "-v" option.

Also, you have a Useless Use Of Cat, and you don't need the extra quotes around the constructed filename.

Code:

#!/bin/bash
.

for ((i=0; i<$detval-1 ; i++)); do

	t="${array[i+1]}"
	awk -v var="$t" '{print $t}' "${array[k]}n${array[i+1]}".d64 > fg.dat 

done

In the future, if you want more detailed help, you really need to supply some sample input data as well for us to work with. It's hard to figure out exactly what's happening from just the code alone.

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/blog/awk-one...ined-part-one/

grail · 10-21-2012, 09:20 AM

One might even hope that you explain what your doing and doing what you explain.

You mention that the code is to cd into a directory, yet I see no evidence in the code presented (explained but not doing)

You also have some mysterious variables 'detval' & 'k' which are never defined or changed so seem pointless (doing but not explained)

And as David aptly pointed out, without any example data it is difficult to understand what you require as you have only shown what does not work.

jkobori · 10-21-2012, 10:11 AM

Here you go:

Code:

#!/bin/bash

grbno=`cat ftest.cat | wc -l`   

read -a y <<< `cat ftest.cat | awk '{print $1}'`
read -a m <<< `cat ftest.cat | awk '{print $2}'`
read -a d <<< `cat ftest.cat | awk '{print $3}'`
read -a s <<< `cat ftest.cat | awk '{print $4}'`
read -a pt <<< `cat ftest.cat | awk '{print $5}'`
read -a at <<< `cat ftest.cat | awk '{print $6}'`



        for ((i=0;i<$grbno;i++))
                do


                                cd /data/fermi/programs/supplementary_programs
                                read -a tomb <<< $(head -1 "${y[$i]}${m[$i]}${d[$i]}${s[$i]}".dat)
                                let detval="${#tomb[@]}"

                                cd /data/fermi/"${y[$i]}${m[$i]}${d[$i]}${s[$i]}"/d64
k=0

                                for ((i=0;i<$detval-1;i++))
                                        do

read -a y <<< `cat /data/fermi/programs/ftest.cat | awk '{print $1}'`
read -a m <<< `cat /data/fermi/programs/ftest.cat | awk '{print $2}'`
read -a d <<< `cat /data/fermi/programs/ftest.cat | awk '{print $3}'`
read -a s <<< `cat /data/fermi/programs/ftest.cat | awk '{print $4}'`
read -a pt <<< `cat /data/fermi/programsftest.cat | awk '{print $5}'`
read -a at <<< `cat /data/fermi/programs/ftest.cat | awk '{print $6}'`

                                                t="${tomb[i+1]}"
                                                cat "${tomb[k]}"n"${tomb[i+1]}".d64 | gawk -v r=$t '{print $(r+1)}' > fg.dat

                                        done

                done

Thank you David the H., I have to use the -v option, now it works!

grail · 10-21-2012, 12:09 PM

So I am curious ... have you just found here strings and want to try them out? Because if not, that looks like a whole lot of hard work.

I won't go through all of it, but just looking at the outer loop and the over use of arrays, you could simply do:

Code:

while read -r y m d s pt at
do
    <your stuff here>
done<ftest.cat

The above replaces:

Code:

grbno=`cat ftest.cat | wc -l`   

read -a y <<< `cat ftest.cat | awk '{print $1}'`
read -a m <<< `cat ftest.cat | awk '{print $2}'`
read -a d <<< `cat ftest.cat | awk '{print $3}'`
read -a s <<< `cat ftest.cat | awk '{print $4}'`
read -a pt <<< `cat ftest.cat | awk '{print $5}'`
read -a at <<< `cat ftest.cat | awk '{print $6}'`



        for ((i=0;i<$grbno;i++)) # this is the outer loop
                do
done

I maybe could have provided more, however, your use of head and multiple changes of directory are a bit confusing. So I leave it with you ... of course if your current method
is working and you are happy then all good