Re: ascii nulls, awk, and tr
From: Richard L. Hamilton (Richard.L.Hamilton_at_mindwarp.smart.net)
Date: 08/03/04
- Next message: all mail refused: "Re: ascii nulls, awk, and tr"
- Previous message: Michael Laajanen: "Re: How do I CLEANLY disconnect a memory stick?"
- In reply to: Gerard C Blais: "ascii nulls, awk, and tr"
- Next in thread: all mail refused: "Re: ascii nulls, awk, and tr"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Tue, 03 Aug 2004 21:22:36 -0000
In article <jppvg0lugiqg7csdiqvtfcqm6fe605j5un@4ax.com>,
Gerard C Blais <gerard.blais@mci.com> writes:
>
>
> I have text files that may contain ascii nulls.
>
> I'd like to delete the null characters.
>
> I also have a lot of other things to do to the rows, and I'm doing
> them in awk already, so adding something to an awk script would be a
> nice way to go.
>
> None of the awks on my Solaris (/usr/bin/awk, /usr/xpg4/bin/awk,
> /usr/bin/nawk) recognize ascii nulls. Someone suggested mawk, and I'm
> trying to get that installed. If anyone knows if mawk will or won't
> work, I'd be interested.
GNU awk (gawk) seems to handle null bytes. If you have the Sun freeware
installed for Solaris 9, it's in /opt/sfw/bin/gawk. If not, depending
on the version of Solaris you have, you might find pre-built binaries at
www.blastwave.org or www.sunfreeware.com. I haven't tried mawk, so I
don't know how well it compares or whether it can handle nulls. But what
I particularly like about gawk is that it seems to handle just about
anything: arbitrarily long lines that neither the supplied awk nor nawk
can handle, nulls just like other characters, etc.
> Meanwhile, it seemed like a perfect fit for tr.
>
> BUT, if I try tr -d \000 null.txt > null.nonull,
>
> tr just hangs, and there is no output. null.txt is a 3 line file
> with nulls in the middle of line 2.
>
> Any ideas?
tr uses standard input, not filenames on the command line (so it was
reading the terminal after you typed that command). Also, older versions
(/usr/bin/tr, /usr/ucb/tr) always strip null bytes, while /usr/xpg4/bin/tr
will handle them just like anything else.
So, either
# changing a space to a space looks like it does nothing, but older
# versions of tr always strip null bytes, so they do the job as a
# side-effect
/usr/bin/tr ' ' ' ' <null.txt >null.nonull
/usr/ucb/tr ' ' ' ' <null.txt >null.nonull
or
# /usr/xpg4/bin/tr handles null bytes like anything else
/usr/xpg4/bin/tr -d '\0' <null.txt >null.nonull
or if you just use "tr", be sure you know which version will be
found first on your PATH and use it accordingly.
-- mailto:rlhamil@smart.net http://www.smart.net/~rlhamil Lasik/PRK theme music: "In the Hall of the Mountain King", from "Peer Gynt"
- Next message: all mail refused: "Re: ascii nulls, awk, and tr"
- Previous message: Michael Laajanen: "Re: How do I CLEANLY disconnect a memory stick?"
- In reply to: Gerard C Blais: "ascii nulls, awk, and tr"
- Next in thread: all mail refused: "Re: ascii nulls, awk, and tr"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|
