Shorten a sed command - using it to make asterisk call logs - sed

Question

I know that asterisk creates it's own call logs in the form of a csv file. For my purposes I need the call logs formatted as I have depicted below. I use:
ls -l /var/spool/asterisk/monitor as the basis for my call logs, which produces this:
-rw------- 1 asterisk asterisk 112684 2013-02-07 17:24 20130207-172424-[phone]-IN.wav
-rw------- 1 asterisk asterisk 44 2013-02-07 17:53 20130207-175311-[phone]-IN.wav
-rw------- 1 asterisk asterisk 2019564 2013-02-07 18:00 20130207-175828-15554561122-OUT.wav
-rw------- 1 asterisk asterisk 44 2013-02-07 22:09 20130207-220805-15554561122-OUT.wav
-rw------- 1 asterisk asterisk 44 2013-02-07 22:12 20130207-221204-15551235566-OUT.wav
-rw------- 1 asterisk asterisk 111084 2013-02-07 22:13 20130207-221255-15551235566-OUT.wav
-rw------- 1 asterisk asterisk 364844 2013-02-07 22:39 20130207-223843-15558271212-OUT.wav
-rw------- 1 asterisk asterisk 4279404 2013-02-07 23:53 20130207-234836-5552785454-OUT.wav
-rw------- 1 asterisk asterisk 44 2013-02-08 00:00 20130208-000026-[phone]-IN.wav
The part I need help with is my command below. It works and produces the exact results I want; however, it seems bulky to me. Can it be shortened?
variables
YESTER=$(date -d "-24 hours" +"%Y-%m-%d-%H%M")
TODAY=$(date +"%Y-%m-%d-%H%M_UTC")
create call log (command I'd like to change)
ls -l /var/spool/asterisk/monitor/ |grep '\.wav'|awk '{print $8 " " $5/1000000}'|sed -e 's/4\.4e\-05/NOT RECORDED/g' -e 's/\.wav//g' -e 's/-/ /g' -e 's/OUT/OUT - Approx Minutes:/g' -e 's/IN/IN - Approx Minutes:/g' -e 's/\(\.[0-9]\).*$/\1/g' -e 's/^.\{15\}/& UTC -/' -e 's/^.\{13\}/&:/' -e 's/^.\{11\}/&:/' -e 's/^.\{6\}/&-/' -e 's/^.\{4\}/& /' -e 's/+//g' > /var/spool/asterisk/monitor/call_logs/${YESTER}__${TODAY}-call-log.txt
For readability here is the command separated by line (without | ):
ls -l /var/spool/asterisk/monitor/
grep '\.wav'
awk '{print $8 " " $5/1000000}'
sed -e 's/4\.4e\-05/NOT RECORDED/g'
-e 's/\.wav//g'
-e 's/-/ /g'
-e 's/OUT/OUT - Approx Minutes:/g'
-e 's/IN/IN - Approx Minutes:/g'
-e 's/\(\.[0-9]\).*$/\1/g'
-e 's/^.\{15\}/& UTC -/'
-e 's/^.\{13\}/&:/'
-e 's/^.\{11\}/&:/'
-e 's/^.\{6\}/&-/'
-e 's/^.\{4\}/& /'
-e 's/+//g'
> /var/spool/asterisk/monitor/call_logs/${YESTER}__${TODAY}-call-log.txt
Output:
2013 02-07 17:24:24 UTC - 15551235566 IN - Approx Minutes: 0.1
2013 02-07 17:53:11 UTC - 15554561122 IN - Approx Minutes: NOT RECORDED
2013 02-07 17:58:28 UTC - 15554561122 OUT - Approx Minutes: 2.0
2013 02-07 22:08:05 UTC - 15554561122 OUT - Approx Minutes: NOT RECORDED
2013 02-07 22:12:04 UTC - 15551235566 OUT - Approx Minutes: NOT RECORDED
2013 02-07 22:12:55 UTC - 15551235566 OUT - Approx Minutes: 0.1
2013 02-07 22:38:43 UTC - 15558271212 OUT - Approx Minutes: 0.3
2013 02-07 23:48:36 UTC - 5552785454 OUT - Approx Minutes: 4.2
2013 02-08 00:00:26 UTC - 15559813232 IN - Approx Minutes: NOT RECORDED

Answer 1

You can put all formating into AWK, why you use sed?
To got it simple, use
[root#gleb monitor]# ls -l --time-style="+%Y-%m-%d %H:%M -"
-rw-r--r-- 1 asterisk asterisk 5195 2013-01-09 21:42 - 20130109-214242-1357756962.1658.WAV
-rw-r--r-- 1 asterisk asterisk 13450 2013-01-13 22:33 - 20130113-223350-1358105630.4124.WAV
Unfortanly i can't give full script, becuase i have other files. Based on such ls command output you now not need rewrite data, so can use it as column.
You can do full features processing including formating in single awk expression.
http://www.gnu.org/software/gawk/manual/html_node/Printf-Examples.html

Answer 2

$ cat tst.awk
{
mins = $5 / 1000000
mins = ( mins == "4.4e-05" ? "NOT RECORDED" : sprintf("%.1f",mins) )
split($8,fname,/-\+?|\./)
date = fname[1]
time = fname[2]
nrs = fname[3]
dir = fname[4]
printf "%s %s-%s ",substr(date,1,4),substr(date,5,2),substr(date,7,2)
printf "%s:%s:%s UTC - ",substr(time,1,2),substr(time,3,2),substr(time,5,2)
printf "%s %s - Approx Minutes: %s\n",nrs,dir,mins
}
$ awk -f tst.awk file
2013 02-07 17:24:24 UTC - 15551235566 IN - Approx Minutes: 0.1
2013 02-07 17:53:11 UTC - 15554561122 IN - Approx Minutes: NOT RECORDED
2013 02-07 17:58:28 UTC - 15554561122 OUT - Approx Minutes: 2.0
2013 02-07 22:08:05 UTC - 15554561122 OUT - Approx Minutes: NOT RECORDED
2013 02-07 22:12:04 UTC - 15551235566 OUT - Approx Minutes: NOT RECORDED
2013 02-07 22:12:55 UTC - 15551235566 OUT - Approx Minutes: 0.1
2013 02-07 22:38:43 UTC - 15558271212 OUT - Approx Minutes: 0.4
2013 02-07 23:48:36 UTC - 5552785454 OUT - Approx Minutes: 4.3
2013 02-08 00:00:26 UTC - 15559813232 IN - Approx Minutes: NOT RECORDED
So just do:
ls -l /var/spool/asterisk/monitor/*.wav | awk -f tst.awk

Answer 3

Although feasible in your case, you should generally avoid parsing ls. I would prefer to see a solution using find instead. If you have access to GNU awk, then you can simplify your pipeline considerably. Run like:
awk -f script.awk <(find /var/spool/asterisk/monitor/ -maxdepth 1 -type f -name "*.wav" -printf "%p %s\n" | sort -n)
Contents of script.awk:
BEGIN {
t = systime()
y = t - 60 * 60 * 24
t = strftime("%Y-%m-%d-%H%M_UTC", t)
y = strftime("%Y-%m-%d-%H%M", y)
}
{
p = "^..(....)(..)(..)-(..)(..)(..)-\\+?([^-]*)-([^\\.]*).*$"
r = "\\1 \\2-\\3 \\4:\\5:\\6 UTC - \\7 \\8 - Approx Minutes:"
s = ($2 != 44 ? sprintf("%.1f", $2/1000000) : "NOT RECORDED")
print gensub(p, r, "", $1) FS s > y "__" t "-call-log.txt"
}
In my testing, this generates a single log file containing your desired output. Because you are now using find, the method could be easily modified if your filesnames were to begin containing whitespace or newline characters. Please let me know how it goes. Cheers.