On Fri, 19 Aug 2005, Florin Iucha wrote: > On Thu, Aug 18, 2005 at 04:22:44PM -0500, Mike Miller wrote: >> On Thu, 18 Aug 2005, Florin Iucha wrote: >>> My script looks for dates as they occur _two spaces_ after some variable >>> piece of text [1]. The awk is more straight forward but can be thrown >>> off by extra logging at the beginning, or spaces in names. >> >> I don't understand how that command can be thrown off. It first egreps >> for the '^From ' lines, so it doesn't get extra lines and it doesn't >> miss lines, then it does the gawk on only those lines. Those lines >> have a uniform format. There are never "spaces in names" because the >> names are always email addresses that cannot have spaces in them. So >> the number of fields is always the same and the gawk will never fail. > > I was thinking about "Florin Iucha" <florin at iucha.net> kind of spaces in > names, but the log message can have (and in my case has) spaces. I was > wrong in the second instance, but right in the first instance. > > Please look again at the snippet I've posted from my .procmailrc . Can that be found here?... http://shadowknight.real-time.com/pipermail/tclug-list/2005-August/author.html It would be good to see someone else's log file I guess. Mine never has a space in an email address because it retains only the address itself, and that obviously cannot contain a space. I have 6.5 years of procmail log file and it contains no examples of extra spaces. Here's the proof: # egrep '^From ' .procmail/log | gawk '{print $3}' | sort | uniq -c 62533 Fri 61566 Mon 35030 Sat 36271 Sun 67671 Thu 66921 Tue 68171 Wed Other people might want to try that to see what their usage patterns look like (and to test for extra spaces). For me, it looks like Wednesday is the biggest email day, then Thursday, then Tuesday, but Monday and Friday get about 8% less than midweek and weekends get about 60% less than weekdays. Have a nice weekend! Mike