Replacing hex characters '\x0D' with '\x0D\x0A'

Shell Programming and Scripting
1

Hello All,

I have a requirement where I need to replaced the hex character - '\x0D' with 2 hex characters - 'x0D' & 'x0A'

I am trying to use SED -

But somehow its not working. Any pointers?

Also the hex character '\x0D' can occur anywhere in the line.

Can this also be accomplished through 'tr'

2

You can use Perl to fix this:

$
$ # create a file that has \x0D characters at multiple places
$ perl -le 'print "\x0D The quick \x0D brown fox jumps \x0D over the lazy dog.\x0D"' >f0
$
$ # check the file contents
$ cat -vET f0
^M The quick ^M brown fox jumps ^M over the lazy dog.^M$
$
$ # do an inline replace of all \x0D characters to \x0D\x0A
$
$ perl -i.bak -pe 's/(\x0D)/$1\x0A/g' f0
$
$ # now check the file contents again
$ cat -vET f0
^M$
 The quick ^M$
 brown fox jumps ^M$
 over the lazy dog.^M$
$
$

tyler_durden

3

you forget to escape \ with \.

sed 's/\\x0D/\\x0D \\x0A/g'

:D:D:D:D

4

Doesn't that remove the special meaning of "\x" which states that what follows is a hex code ?

$
$ printf "abc\x0Ddef" | od -bc
0000000 141 142 143 015 144 145 146
          a   b   c  \r   d   e   f
0000007
$
$ # sed
$ printf "abc\x0Ddef" | sed 's/\\x0D/\\x0D\\x0A/g' | od -bc
0000000 141 142 143 015 144 145 146
          a   b   c  \r   d   e   f
0000007
$
$ # perl
$ printf "abc\x0Ddef" | perl -ne 's/\x0D/\x0D\x0A/g; print' | od -bc
0000000 141 142 143 015 012 144 145 146
          a   b   c  \r  \n   d   e   f
0000010
$

tyler_durden

5

Thanks 'durden_tyler' your code worked...but just to mention that it didn't work on Windows...

Same was the case with my 'sed' command I had mentioned at the start. It was not giving the desired results on Windoze but when I executed on Linux it worked like a charm...

I think its due to the fact that interpreters like 'sed', 'Perl' ignores 'CR' (chariage return) characters on windoze..........

6

Don't know about "sed" on Windows, but that would be highly unusual for Perl on Windows:

C:\>
C:\>REM create a file that has \x0D characters at multiple places
C:\>perl -le "print \"\x0D The quick \x0D brown fox jumps \x0D over the lazy dog.\x0D\"" >f0
C:\>
C:\>REM check the file contents
C:\>vis <f0
\015 The quick \015 brown fox jumps \015 over the lazy dog.\015
C:\>
C:\>REM do an inline replace of all \x0D characters to \x0D\x0A
C:\>perl -i.bak -pe "s/(\x0D)/$1\x0A/g" f0
C:\>
C:\>REM now check the file contents again
C:\>vis <f0
\015
 The quick \015
 brown fox jumps \015
 over the lazy dog.\015

C:\>
C:\>REM print file using the DOS "type" command
C:\>type f0
 The quick
 brown fox jumps
 over the lazy dog.

C:\>
C:\>

The perl interpreter used in the example above is from the ActiveState Perl installed using MSI.

"vis" is just a small C program that makes non-printable characters visible in octal code.

tyler_durden

7

paragkalra,

Without using Perl or C, may be this is what you were looking for,
a pure sed solution, on a typical Unix environment:

  sed 's/\\r/\\r\\n/g' &lt;infile &gt;outfile

At least it works flawlessly on my RH/Fedora... I've tested it using
an input file, as shown by "od -c", like:

0000000 f i r s t l i n e \r s e c o n
0000020 d l i n e \r t h i r d l i n
0000040 e \r
0000042

and having the output file generated like this:

0000000 f i r s t l i n e \r \n s e c o
0000020 n d l i n e \r \n t h i r d l
0000040 i n e \r \n
0000045

Please not that \r and \n are the standard escape sequences
(used by both "od" and "sed") for 0x0D and 0x0A - presuming
we live in an ASCII world, of course. This may not apply if
you are dealing with systems built upon different codings...

Helpful?

Mario.