Originally Posted by agama

Using bash is perfectly acceptable. Kshell has a few features that make it more attractive to me, but that's a personal preference.

The 2>&1 causes output to stderr (file descriptor 2) to be written onto stdout (file descriptor 1) which is what the shell maps to the process on the other side of the pipe.

Some comments on the awk:

Code:

awk ' 
$1+0 > 0 {   # when the first field converted to a number is greater than 0
   n = $1+0; # save the first field as a number in variable n
   next;      # continue processing with next line of input (not required for this case, but good form)
} 

END {     # execute this code after the last line of input has been processed
   print n;  # print the value of n to standard output
}'

In short, awk reads the input file (or from the pipe a.k.a. stdin) and for each line of input executes the programme. In this case the programme is a very simple check to see if the first column (or field if you like that phrase) contains a digit, and if it does the code in curly braces is executed (saving that value in n).

It uses a bit of awk trickiness to do this. The field is added to 0 knowing that awk will convert any text string to 0 and for each record that does not start with a number (e.g. 1.) the result will be 0 + 0 and the condition will be false; the code will not be executed and n will not be changed.

Here's a link to a decent tutorial on awk -- worth the read as awk is one of the (IMHO) most useful tools available to a UNIX programmer.

You are only allowed to post URLs once you have at least 5 posts.

Hope this helps.