Substitute variable inside nawk

Shell Programming and Scripting
1

Hi,

I need to set "prd" in the below command to a unix variable

nawk '/^#/ {next} FNR==NR {prd[$0];next} !($0 in prd)'

So, this is what i did

fname=prd   // unix shell variable

nawk -v fname=$fname '/^#/ {next} FNR==NR {fname[$0];next} !($0 in fname)'

But the value of fname i.e "prd" is not getting substituted in the command.

Can you tell me why ?

OS: SunOS mymac 5.10 Generic_150400-09 sun4v sparc SUNW,SPARC-Enterprise-T5220
2

You have 2 variables called 'fname':

  1. Scalar fname passed into awk by: -v fname=$fname
  2. Array fname you're populating with $0: fname[$0]

What's the purpose of passing a scalar fname to awk?
What exactly are you after?

3

As i explained in the OP all i want is instead of using "prd" in the command i wish to use unix variable fname.

So, the problem is the below is not working !!

  1. Scalar fname passed into awk by: -v fname=$fname

Can you now help ?

4

Maybe this will show you some light

echo 'Show me the content of $fname' | awk -v fname=$fname '{print fname}'

If you can see prd in your stdout, then that's not the problem.

5

Let me rephrase my question.

I am using the below code to find the difference between two files prod.properties and back.properties.

nawk '/^#/ {next} NR==FNR {prod[$0]=$0;next} {if ($0 in prod) { delete prod[$0] } else {print }}; END {for (i in prod) print prod}' prod.properties back.properties

Source: Comparing two files issue. Post: 302903183

Now, the requirement is to automate the above code so that it does the job for any two file names not neccessarily prod.properties and back.properties

I am passing any two filenames like this.

./check.sh filename1.properties filename2.txt

and reading them in variables file1=$1, file2=$2.

So, how can I achieve that ?

6

Assuming that check.sh contains:

nawk '/^#/ {next} NR==FNR {prod[$0]=$0;next} {if ($0 in prod) { delete prod[$0] } else {print }}; END {for (i in prod) print prod}' prod.properties back.properties

change it to:

file1="$1"
file2="$2"
nawk '/^#/ {next} NR==FNR {prod[$0]=$0;next} {if ($0 in prod) { delete prod[$0] } else {print }}; END {for (i in prod) print prod}' "$file1" "$file2"

or just:

nawk '/^#/ {next} NR==FNR {prod[$0]=$0;next} {if ($0 in prod) { delete prod[$0] } else {print }}; END {for (i in prod) print prod}' "$1" "$2"

or even:

nawk '/^#/ {next} NR==FNR {prod[$0]=$0;next} {if ($0 in prod) { delete prod[$0] } else {print }}; END {for (i in prod) print prod}' "$@"
7

So the "prod" inside nawk is just a variable ?

I thought it has some relation with the prod.properties

Please confirm.

8

prod is just an array in your script, there is no relation with file prod.properties

9

Yes, you are right. That array name within awk is a variable that you can call what you like. It's not bad an idea to use some mnemonic name indicating what it's used for, for e.g. later debugging.