On Sun, 2009-01-04 at 13:48 -0500, Victor Duchovni wrote: > On Sun, Jan 04, 2009 at 08:33:12PM +0530, Ramprasad wrote: > > > My logfiles are approx 1.5 GB each, if we rotate twice a day > > grep sender.*recipient $logfile is really much much faster than grep > > -i sender.*recipient $logfile > > IMHO grep would not count as lousy code ( on linux ) > > Use "egrep -i", and make sure you have "LANG=C" in the environment (or > equivalent). Doing case-insensitive Unicode matches is likely much more > painful than ASCII matches. >
Thanks that made *a lot* of difference with "env LANG=C" grep -i works fine. Still I think a log in lowercase would be a nice feature to have. We also put the entries into a DB and generate statistics. Every logparser script has to take care to convert to lowercase. Even if performance is not an issue, it would be a lot neater to log in lowercase
