Using awk's system() command to process part of a file
------------------------------------------------------

Example: I want to use awk to select PART of a file for processing, 
but I want to use a different command (like sed or fmt) to do the 
work. Here's how to do it against myfile.

                                        
awk '$6 > 4 {system("echo " $0 "| sed -f script.sed");next};1' myfile

or more succinctly:

     awk -f myscript.awk myfile

where my myscript.awk is:

     #---begin  myscript.awk ---
     $6 > 4 {
       system ("echo " $0 "| sed -f script.sed")
       next
     }
     1
     #--- end of script ---

Every line will be printed, but each line in which the value of the
6th field ($6) is greater than 4 will be passed to sed for processing.

Note that the space after 'echo' is significant: "echo ".

The '1' just before the end of the script tells awk to print all the 
lines normally that were not matched by the pattern. Technically, '1' 
is a pattern itself that resolves to TRUE, and if a pattern is 
specified in awk without an action, the line will be printed when the 
pattern matches or is true. Thus, '1' works to print every line.

Finally, note that this is a line-oriented script. To pass a group of 
lines to an external program, collect them in a variable and echo the 
variable, piping the results to the program you specify.

[eof]