[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: grep --ignore-case efficiency
From: |
Paulo Nogueira |
Subject: |
Re: grep --ignore-case efficiency |
Date: |
Wed, 30 Jan 2013 18:44:00 +0000 (WET) |
User-agent: |
Alpine 2.00 (LFD 1167 2008-08-23) |
Thank you Arnold and Johannes for your comments.
For convenience I'll do a mixed reply below.
The simple solution was to do
export LC_ALL=C
and then run grep, which would give you a big speedup.
Tried it and, indeed, this gives a big speedup!
Comparable times on an old version of grep (2.5.1),
but on a recent installation of Fedora (grep 2.14)
it does not work just as well because it's still about
10x slower (in my test case, with a simple text string).
Ten times is still better than ninety times, although
it's not expected after seing the result of grep 2.5
Have a look at
http://en.opensuse.org/SDB:Plain_Text_versus_Locale
in particular the section
"Processing UTF-8 is slower than ISO-8859 and ASCII"
and replace the "wc" tool therein with "grep".
Tried also
export LC_ALL=POSIX ; export LANG=POSIX
as explained in the web page you cited, but still no
improvements on Fedora 18. Maybe there is some other
factor to take into account.
Thanks,
Paulo