topic Stickey NUL issue ... sed/grep/awk ... perl in Operating System - Linux

Stickey NUL issue ... sed/grep/awk ... perl

A. Daniel King_1 — Tue, 12 Apr 2005 07:40:59 GMT

Hi, folks,

How do I (1) detect and (2) replace a mid-line null, such as:

This is a li^@ne in a text file ...

Where the ^@ is a NUL character?

Re: Stickey NUL issue ... sed/grep/awk ... perl

RAC_1 — Tue, 12 Apr 2005 08:00:28 GMT

Post following.
cat "file" | od -c
cat file | vis

Would like to look at how it looks like.

Re: Stickey NUL issue ... sed/grep/awk ... perl

A. Daniel King_1 — Tue, 12 Apr 2005 08:03:48 GMT

cat /tmp/1.txt | od -c
0000000 T h i s i s \0 t h e f i l e
0000020 .
0000021

cat /tmp/1.txt | vis
This is\000the file.

Re: Stickey NUL issue ... sed/grep/awk ... perl

Kent Ostby — Tue, 12 Apr 2005 08:08:53 GMT

In awk, I think you could do:

{idx1=index($0,"\0");
if (idx1==0) {print $0;next}
newstring=substr($0,1,index-1)substr($0,index+1);
}

Re: Stickey NUL issue ... sed/grep/awk ... perl

RAC_1 — Tue, 12 Apr 2005 08:08:56 GMT

cat file | tr "\000" "\xxx"

xxx is ascii char you want to replace 000 with. man ascii for details.

Anil

Re: Stickey NUL issue ... sed/grep/awk ... perl

Kent Ostby — Tue, 12 Apr 2005 08:09:28 GMT

Let's try that again:

In awk, I think you could do:

{idx1=index($0,"\0");
if (idx1==0) {print $0;next}
newstring=substr($0,1,index-1)substr($0,index+1); print newstring;
}

Re: Stickey NUL issue ... sed/grep/awk ... perl

Hein van den Heuvel — Tue, 12 Apr 2005 09:46:06 GMT

I like tr -d best myself, but it all works:

# perl -e 'printf "Thisis%cthefile\n",0' > x
# cat x
Thisisthefile
# od -c x
0000000 T h i s i s \0 t h e f i l e \n
#
# perl -pe 's/\0//' x | od -c
0000000 T h i s i s t h e f i l e \n
#
# cat x | tr -d "\000" | od -c
0000000 T h i s i s t h e f i l e \n
#
# awk '{gsub(/\000/,""); print}' x | od -c
0000000 T h i s i s t h e f i l e \n

Hein.

Re: Stickey NUL issue ... sed/grep/awk ... perl

A. Daniel King_1 — Mon, 22 May 2006 13:40:42 GMT

awk had trouble, I ended up using perl, and I really like the vis and tr solutions.