NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/unix/question/16005

Internet Message Format | 1993-01-28 | 3.7 KB

Path: sparky!uunet!ferkel.ucsb.edu!taco!gatech!paladin.american.edu!howland.reston.ans.net!spool.mu.edu!olivea!sgigate!odin!akshay.csd.sgi.com!vidya
From: [e-mail] (Vidya Alankar)
Newsgroups: comp.unix.questions
Subject: Re: Getting the nth line of a file
Message-ID: <[e-mail]>
Date: 26 Jan 93 23:27:08 GMT
References: <[e-mail]> <[e-mail]>
Sender: [e-mail] (Net News)
Organization: Silicon Graphics, Inc.
Lines: 103
Nntp-Posting-Host: akshay.csd.sgi.com

In article <[e-mail]>, [e-mail] writes:
|> I wrote:
|> : Is there an easy, elegant, and quick way to get the Nth line of a file in
|> : UNIX?  Or, failing that, in C?
|>
|> : Currently I use the simple, but somewhat slow:
|>
|> : head -n | tail -1
|>
|> I received several email replies, which I will summarize here.
|>
|>
|> [e-mail] (Rouben Rostamian)
|>
|> > To list the sixth line of a file, use:
|>
|> > awk '{NR==6 {print; exit}' filename

awk '{ if ( NR == 6 ) { print; exit } }' filename

|>
|>
|> Unfortunately, I couldn't get this one to work.  But then, I'm not familiar
|> with awk.  Perhaps people who are familiar with it can correct the above.
|>
|>
|> Mark Borges <[e-mail]>
|>
|> > I think the following may be faster and is equivalent:
|>
|> > sed -n 'N p' file
|>
|> > where -n supresses echoing the file to stdout, N is the Nth line you
|> > want, p means to print it, and file is the input file. I'm not sure if
|> > sed quits after printing the Nth line, which is really what you want
|> > for speed; the sed man page always confuses me.

It does read the full file. I do not know enough about sed to say if they
is a way to stop after the Nth line. The awk command above does just this.

|>
|> > Hope this works fast enough.
|> > Mark
|>
|> This one works just fine, and fast enough for my purposes.
|>
|>
|> Thayne Forbes <[e-mail]>
|>
|> > Ack! Barf!
|> > Try this.
|>
|> >     FLINE=`sed -n "$COUNT p" <$FILTER`
|> > This is from my news filter program.  COUNT is the line number, and FILTER
|> > is the file I am reading.  This reads the one line and assignes it to
|> > FLINE.
|>
|> > Remember, sed is your fiend!
|>
|> Yes, I know that sed is my friend, and one day I'll have to get into it more
|> seriously.  The above also works.
|>
|>
|> ptsfa![e-mail]
|>
|> > Try sed(1):
|>
|> >     sed -n Np
|>
|> > where N is the desired line number.
|>
|> > For example, the third line of /etc/group would be:
|>
|> >     sed -n 3p /etc/group
|>
|> > Dave Turner    [phone]    {att,bellcore,sun,ames,decwrl}!pacbell!dmturne
|>
|> This looks the most promising, being a direct statement with no quotes and no
|> redirection, which fits my definition of elegant.  (So sue me, I want it to
|> look nice as well as work.)
|>
|>
|> Chris Sherman <[e-mail]>
|>
|> > Maybe:
|>
|> > sed -n '30 {p;}' < /etc/passwd
|>
|> > Will show the 30th line in /etc/passwd.
|>
|> This also works.
|>
|> So it seems that sed is the way to go, and some combination of the line number
|> followed by p.  Since I'm calling this from a C program where I write to a
|> string and then send it to system(3), it's convenient to use Dave Turner's
|> method, though I imagine others might work better in scripts.  Thank you all
|> for your help.
|>
|> --
|> ---Alfvaen(Still looking for "October's Baby")
|> "Clocks don't bring tomorrow--knives don't bring good news."                       ---Bruce Cockburn
|> Current Album--Mike Oldfield:The Killing Fields Soundtrack
|> Current Read--Sean Russell:The Initiate Brother