Quoting Clive Standbridge on 2010-01-26 15:58:00: > Brian seems to have thought of everything. Not everything :) Trying to learn, to the point of mastery, two languages at once though. Eric's (no pipe to wc) works far better when you're just interested in the number of lines that match, and not interested in the count -- and runs faster to boot. As an example, given the file ...  foobar foobaz foovax kremvax vaxvax super-foonly foo fighters [end 1] If we wonder only how many lines have 'foo' in them, and we use Eric's construct, we get the following: bryans@esterhazy:~$ do strings /tmp/foo | grep -oce foo # prints 3 Slightly different tools for slightly different problems, I guess :) Now, running each 1000 times on my 1.1GHz P3 laptop: bryans@esterhazy:~$ time for x in $(seq 1000); do strings /tmp/foo | grep -oce foo >/dev/null; done # How many lines contain 'foo'? bryans@esterhazy:~$ time for x in $(seq 1000); do strings /tmp/foo | grep -oe foo | wc -l >/dev/null; done # How many times does 'foo' occur? Eric Brian real 8.675 12.172 user 2.948 4.008 sys 5.340 7.800 This thread actually gave me a slightly better understanding of grep and pipelines. I assumed that adding a pipe would slow it down a bit, but never could get off my corpulent posterior to look into put some hard data into my hyptothesis. -- _ Brian Ryans 8B2A 54C4 E275 8CFD 8A7D 5D0B 0AD0 B014 C112 13D0 . ( ) ICQ UIN: 43190205 | Mail/MSN/Jabber: BrianLRyans@gmail.com ..: X ASCII ribbon campaign against HTML mail and v-cards: asciiribbon.org / \ This is my .signature. There are many like it, but this one is mine.
Description: Digital signature