I like Common LISP and I also like awk. Dealing with text files in Common LISP
is often painful. So I wrote a small awk like common lisp macro, which helps a
lot dealing with text files.
Here is the implementation, I used the uiop package for split-string function,
it comes with sbcl. But it's possible to write your own split-string or reused
the infamous split-str function shared on the Internet.
(defmacro awk(file separator &body code)
"allow running code for each line of a text file,
giving access to NF and NR variables, and also to
fields list containing fields, and line containing $0"
`(progn
(let ((stream (open ,file :if-does-not-exist nil)))
(when stream
(loop for line = (read-line stream nil)
counting t into NR
while line do
(let* ((fields (uiop:split-string line :separator ,separator))
(NF (length fields)))
,@code))))))
It's interesting that the "do" in the loop could be replaced with a "collect",
allowing to reuse awk output as a list into another function, a quick example I
have in mind is this:
;; equivalent of awk '{ print NF }' file | sort | uniq
;; for counting how many differents fields long line we have
(uniq (sort (awk "file" " " NF)))
Now, here are a few examples of usage of this macro, I've written the original
awk command in the comments in comparison:
;; numbering lines of a text file with NR
;; awk '{ print NR": "$0 }' file.txt
;;
(awk "file.txt" " "
(format t "~a: ~a~%" NR line))
;; display NF-1 field (yes it's -2 in the example because -1 is last field in the list)
;; awk -F ';' '{ print NF-1 }' file.csv
;;
(awk "file.csv" ";"
(print (nth (- NF 2) fields)))
;; filtering lines (like grep)
;; awk '/unbound/ { print }' /var/log/messages
;;
(awk "/var/log/messages" " "
(when (search "unbound" line)
(print line)))
;; printing 4nth field
;; awk -F ';' '{ print $4 }' data.csv
;;
(awk "data.csv" ";"
(print (nth 4 fields)))