cat|cut|cut|sort|uniq|sort > file

[David O'Meara]

So it turns out running this on 24Gb takes a while.

Who knew?

[David O'Meara]

actually I've been running the following on the combined log files for the month from 100 client sites. Personally I think the problem is too much logging.

cat *.log* |cut -d' ' -f7|cut -d'?' -f1|sort |uniq -c|sort -n>file.count

[David O'Meara]

Ho hum.

I was hoping to have a look at the results, but I might go to bed.

[Doug Slattery]

Maybe this will save you a second or two . (Warning, I didn't test it, but I think you get the idea...)

sed -e 's/^\([^ ]* \)\{6\}\([^?]*\)?/\2/' < *.log* > reallybig.log; sort -n reallybig.log | uniq -c > file.count

Aloha,
Doug

-- Nothing is impossible if I'mPossible

[Pat Farrell]

When N is big, even N ln(N) is big

[Doug Slattery]

yes, but at least now, it's not 2N

Aloha,
Doug

-- Nothing is impossible if I'mPossible

[David O'Meara]

keeping in mind we're in MD and not Linux, I'll just say "One day I need to look at sed"