-----------------------------------------------------------------------------
Using Awk Arguments Directly

During the BEGIN{..} code block, ARGV has not been looked at, but has
been read into an array.  As such you can look through it an act or
modify on those variables.

  =======8<--------
  awk 'BEGIN { print "data = " data
               for (i=0;i<ARGC;i++) print "argv[" i "] = " ARGV[i]
       }' data=2 abc

  data =
  argv[0] = awk
  argv[1] = data=2
  argv[2] = abc
  =======8<--------

Note data has not been set in the BEGIN block. Even though the argument array
has been defined in memory.   Variables assigned using a -v option however
is available to the BEGIN block.

  awk -v data=3 'BEGIN { print "data = " data }'
  data = 3

If you don't want to process some file (or modify what file to process),
you can modify the ARGV array in the BEGIN block.   For example
extract an element and remove it from the file list to be processed...

  =======8<--------
  #!/bin/nawk -f
  BEGIN { awkoption = ARGV[1]; delete ARGV[1] }
  ...
  =======8<--------

NOTE: if delete ARGV[1] fails try ARGV[1]="" or "/dev/null"


Solaris 9:
   "awk" attempts to open all given files before BEGIN (error if not present)
   Also data assignment has not been done before BEGIN!
   "awk" also does not have access to -v option or to "ARGV" array.
   However "nawk" works as documented above!

-----------------------------------------------------------------------------
Awk processing two files differently

This could be used to read a config or a table of data before the data
to awk.

Variable settings take effect when they are encountered on the command line,
so, for example, you could instruct awk on how to behave for different files
using this technique.  For example:

  awk 'data==0 { print "process config_data here"  }
       data==1 { print "process data_file here" }
      ' config_data data=1 data_file

Note that some versions of awk will cause variable settings encountered
before any real filenames to take effect before the BEGIN block is
executed, but some won't so neither way should be relied upon.

Solaris 9: ou must use "nawk" for this to work.

-----------------------------------------------------------------------------
Variables in Regular Expressions

You cannot use a variable inside a regular expression.
But you can use the whole variable as the regular expresion.

 awk '$0 ~ RegExp  { print $0 }' RegExp="sock" /usr/share/dict/words


-----------------------------------------------------------------------------
Multi-dimentional Arrays are possible in awk

  # found that bitmap[$0] would not work.
  # so assign to variables.
  coordx = $1+0
  coordy = $2+0
  # assign "X" to that element of array
  bitmap[coordx,coordy] = "X"

-----------------------------------------------------------------------------
Reading from commands in awk

The trick is to CLOSE the file that you use for the command pipe

    "date" | getline x;   close "date";

-----------------------------------------------------------------------------
Multiple Value returns from awk...

To do this have awk output a shell script. This can then be evaluated!
And you can generate as many variables as you like. (C-programs do this too)

eval `awk 'END { print "FLAG="(NR > 0) ? 1 : 0 }' -`

echo "FLAG=$FLAG"

-----------------------------------------------------------------------------
Awk field separator

The newer gnu-awk allows you to specify multi character and regular expresion
field seperators.

  awk -F"[\"><]+" '....'

-----------------------------------------------------------------------------
Changing the record seperator inside an awk script.

Awk reads and record separates its input before the script properly starts
as such you need to get it to re-evaluate that input.

For example...
    awk 'BEGIN { RS=";" }
         { $1=$1; printf "%s%1s\n",$0,";" }
        ' infile

The $1=$1 is required to get awk to re-compute the $0 for the first record.

-----------------------------------------------------------------------------
Determine which column numbers of relevant 'ps' fields

  =======8<--------CUT HERE----------axes/crowbars permitted---------------
  #!/bin/sh
  CMDCOL=`ps -e | awk '
          NR == 1 { for (i = 1; i <= NF; i++)
                      if ($i == "COMMAND" || $i == "CMD" || $i == "COMD")
                        cmdcol = i
                  }
          END     { print cmdcol }
  '`

  if [ "$CMDCOL" = "" ]
  then
          echo "$0: Unrecognised ps format for COMMAND field"
          exit 1
  fi

  TTYCOL=`ps -e | awk '
          NR == 1 { for (i = 1; i <= NF; i++)
                      if ($i == "TTY")
                        ttycol = i
                  }
          END     { print ttycol }
  '`

  if [ "X$TTYCOL" = "X" ]; then
      echo "$0: Unrecognised ps format for TTY field"
      exit 1
  fi

  #
  # Print list of all terminals running program
  #
  echo -n "Terminals running vim : "
  ps -e | awk ' $'"$CMDCOL"' == "vim" { print $'"$TTYCOL"' }'

  =======8<--------CUT HERE----------axes/crowbars permitted---------------

-----------------------------------------------------------------------------