49.7 Total a Column with addup

addup	Some programs output information in columns. The addup script totals the numbers in a column. It reads from files or standard input. For example, the lastcomm command shows CPU time used in column 4, like this:

% lastcomm tcomm


sleep            tcomm    __         0.08 secs Thu Mar 27 10:23
date             tcomm    __         0.08 secs Thu Mar 27 10:23
tail             tcomm    __         0.09 secs Thu Mar 27 10:23
pwho             tcomm    __         0.30 secs Thu Mar 27 10:23
% lastcomm tcomm | addup 4


0.550000

grep -c (15.8 ) outputs the number of matches after a colon (: ) on each line. To total the matches, pipe grep 's output through a little sed command to strip off the filenames and colon; have addup sum the output (the "first column"):

% grep -c CAUTION *.txt


abar.txt:0
applic.txt:3
badprob.txt:235
   ...
% grep -c CAUTION *.txt | sed 's/.*://' | addup 1


317

Here's the script:

case "$1" in
[1-9]*) colnum="$1"; shift;;
*) echo "Usage: `basename $0` colnum [files]" 1>&2; exit 1;;
esac

# Use integer output, but switch to %.4f format if "." in input.
awk '{sum += $col}
END {print sum}' col=$colnum OFMT='%.4f' ${1+"$@"}

The ${1+"$@"} holds filenames (if any) from the command line and works around a shell quoting problem (46.7 ) . The awk script passes in the column through a variable on its command line, then $col becomes the column number. The script defaults to integer output format, without a decimal point. If it sees a "." in the input (like 1.5 ), it switches to floating-point output format; the OFMT variable setting of %.4f forces awk to always print the result with four digits after the decimal point. (The default OFMT setting, %.6g , prints large numbers in e-notation. If you want that, delete the OFMT='%.4f' .)