How to get a count of variables passed to awk

Question

I am trying to invoke awk inside a bash script and passing some bash variables values to it. Is there any way to get a count of these variables and print all which are passed to it inside awk.

No. of variables passed to awk would be dynamic.

Below example

#!/bin/bash
NAME="Dell"
SERIAL="12345"
echo "Hello" | awk -v a=$NAME -v b=$SERIAL '{ print a,b }'

Referenced, Working example, Irvur's answer in here

#!/bin/bash

function confirm_peers (){
  header_def=("${header_def[@]:-None}")
  read -r -p "Enter nodes from header[${header_def[*]}]: " -a header
  header=("${header[@]:-${header_def[@]}}") ;
  echo -e "\nInput Recorded - ${header[@]} \n"
}

createEncConfig()
{
 /usr/bin/awk -f - <<-'EOE' ${header[*]}  FS=":" $1
  BEGIN {
    count=1
    for (i = 1; i < ARGC - 2; ++i) {
    split(ARGV[i], ar, "=");
    print ar[2]
    }
print "\nCount=" ARGC -3
        }
EOE
}
confirm_peers
# Making array content adaptable to awk input
for x in ${!header[@]} ; do header[x]="a=${header[x]}"; done
createEncConfig $1

Referenced - Easiest so far , Kusalananda's answer

#!/bin/bash

function confirm_peers (){
  header_def=("${header_def[@]:-None}")
  read -r -p "Enter nodes from header[${header_def[*]}]: " -a header
  header=("${header[@]:-${header_def[@]}}") ;
  echo -e "\nInput Recorded - ${header[@]} \n"
}

createEncConfig()
{
/usr/bin/awk -v args="${header[*]}" -f - <<-'EOE' FS=":" $1
BEGIN {
  count=split(args,ar," ")
  for ( x in ar ) {
    print ar[x]
  }
print "\n" count
}
EOE
}
confirm_peers
createEncConfig $1

Output : Just pass any dummy file

$ bash /tmp/a.bsh /tmp/enc1.txt

Enter nodes from header[None]: a b c d

Input Recorded - a b c d

a
b
c
d

Count=4

Thanks to all ..

Answer 1

One option I can think of is to reposition those variables so they feature in ARGV. ARGC will indicate how many of them there are. Beware this makes the variables unavailable in BEGIN block. You also need to account for any file names that you may be passing as parameters to awk by subtracting the number of such files from ARGC

echo "Hello" | awk  '{ print a, b, ARGC - 1}' a=$NAME b=$SERIAL
DELL 12345 2

Answer 2

From your comment, it seems that you have an array in bash that you'd like to pass into an awk program on the command line.

Here's an easy way of doing that:

array=("my 1st item" "my 2nd item" "the last item")

(
    IFS=:

    awk -v var="${array[*]}" '
        BEGIN {
            array_len = split(var, array, ":")
            for (i = 1; i <= array_len; ++i)
                printf("Item #%d is \"%s\"\n", i, array[i])
        }'
)

This would output

Item #1 is "my 1st item"
Item #2 is "my 2nd item"
Item #3 is "the last item"

This takes the array and turns it into a single string of :-delimited items. If any of the elements of the array contains :, then pick another character for delimiting the items in the string. The expansion of ${array[*]} will join the items in the array on the first character of $IFS, which is why we set this variable before performing that expansion. This string is passed to awk as the variable var.

In the BEGIN block in the awk program, we then create the array array and also keep track of the array's length in array_len through splitting var on :. The loop is just for showing that we managed to split the string successfully.

We set IFS and run our awk program in a subshell so that the rest of the script can run with an unmodified IFS variable.

Answer 3

Now, as awk doesn't provide means to get at its own command line and the options therein, this is a bit far fetched, and it relies on linux internals like the proc file system, but just for the fun of it (YMMV):

awk -v"VAR1=sd" -va=b -vJ=onny '
BEGIN   {("tr \47\\000\\n\47 \47\t\47 </proc/$PPID/cmdline") | getline CMDLN
         n = split (CMDLN, ELM, "\t")
         for (i=1; i<=n; i++)   if (sub ("^-v", "", ELM[i]))    {split (ELM[i], T, "=")
                                                                 print "Option", T[1], "=", T[2]
                                                                 Options++
                                                                } 
        }
'
Option VAR1 = sd
Option a = b
Option J = onny

The "command" being read by getline will be executed in a subshell, so we get at the cmdline of the parent PID, and tr the <NUL> and <LF> chars to <TAB>s. Then, the CMDLN is split into its elements, in each of which a leading "-v" is matched (and eliminated) to find variable definitions. Those are printed, and the options count incremented. Give it a try.

Answer 4

You can pass your array into awk directly with here-string.

awk -F'SEP' '{ print NF }' <<<"${header[*]}"

Where the SEP is the character which your array elements are delimited with.

Tset:

$ array=( 1, 2, 3, 7 )
$ awk -F',' '{ print NF }' <<<"${array[*]}"
4

And $1 will be your first element's value, $2 second and etc.

Answer 5

Since you have a bash array, and bash knows how many elements are in the array, simply tell awk directly:

$ array=("my 1st item" "my 2nd item" "the last item")
$ awk -v var="${array[*]}" -v nargs=${#array[@]} '
  BEGIN {
    print nargs;
  }
'
3