Using sed to replace text with carriage return and line feed

Question

Using sed to replace text with carriage return and line feed

HI,
Im trying to do a search and replace on a text file using sed. I'm trying to replace the string "000W" with \r\n.

I've tryed:
sed 's/000W/\r\n/g' test.prn

Now wherever there is the string "000W" it is replaced by rn and not \r\n which is what I expected.

e.g.
echo "test1;test2;test3;test4" | sed 's/;/\r\n/g'
produces
test1rntest2rntest3rntest4

Any clues?

Jeff.

MacBook, Mac OS X (10.5.7)

Posted on Jun 4, 2010 1:08 AM

Reply

Answer 1

Best reply

BobHarris

Level 9

53,291 points

Jun 4, 2010 7:09 AM in response to Jeffrey Young

Try this instead:


echo "test1;test2;test3;test4" | perl -ne 's/;/ /g; print' | cat -vte
test1^M$
test2^M$
test3^M$
test4$

Where the 'cat -vte' display ^M for the carriage return, and $ for the newline.

So for your real goal, this should work:


perl -ne 's/000W/ /g; print' test.prn

Reply

Answer 2

Tony T1

Level 6

10,232 points

Jun 4, 2010 5:04 PM in response to Jeffrey Young

This will work:


echo "test1;test2;test3;test4" | sed 's/;/'$' ''
/g'

Two issues
1) in sed, "However, if you are inserting a new line, don't use " " - instead insert a literal new line character" (see: http://www.grymoire.com/Unix/Sed.html section "Using newlines in sed scripts"
2) in bash, to insert a carriage return, use $' '


Words of the form $'string' are treated specially. The word expands to
 string, with backslash-escaped characters replaced as specified by the
 ANSI C standard. Backslash escape sequences, if present, are decoded
 as follows:
 a alert (bell)
   backspace
 e an escape character
   form feed
 new line
 carriage return
       horizontal tab
 v vertical tab
 \ backslash
 \' single quote
 nn the eight-bit character whose value is the octal value
 nnn (one to three digits)
 xHH the eight-bit character whose value is the hexadecimal
 value HH (one or two hex digits)
 cx a control-x character
 The expanded result is single-quoted, as if the dollar sign had not
 been present.
 A double-quoted string preceded by a dollar sign ($) will cause the
 string to be translated according to the current locale. If the cur-
 rent locale is C or POSIX, the dollar sign is ignored. If the string
 is translated and replaced, the replacement is double-quoted.

Reply

Answer 3

Jun 4, 2010 5:56 PM in response to BobHarris

Thanks, that worked.

I now need to find out how to search for multiple "0.000E+00"s in my file elegantly using perl.

Reply

Answer 4

BobHarris

Level 9

53,291 points

Jun 4, 2010 9:00 PM in response to Jeffrey Young

I now need to find out how to search for multiple "0.000E+00"s in my file elegantly using perl.

What do you want to do with the 0.000E+00 values? Change them to carriage return, line feeds?

The perl -n option says to apply the perl script to each line of the input. The perl -e '...' says the '...' following the -e is a script to be executed.


perl -ne 's/0.000E+00/ /g; print' input.file

But if you want to do something else with those 0.000E+00 values, we are going to need some additional guidance.

Reply

Answer 5

Jun 7, 2010 5:07 AM in response to BobHarris

I ended up reading a few man pages on Perl. I like it. I've written my first perl script to deal with converting a dump file (from a piece of equipment) to something MathCAD can import. I needed to discard the 0.000E+00 numbers plus a few other things. There were a few things that bailed me up but I found out why, and I now have a script file that executes blazingly fast - I really like this.
Essentially three lines of perl script do the job after the end of the header is found:
.
.
.
for ($hdrline=<$in>; $hdrline!~/000W/; $hdrline=<$in>){} # Positions the file pointer after the header
.
.
s/\r\n/ /; # Replace <CR><LF> with ' '
s/000W\sZR\d\s(\d s){4}/\r\n/; # Replace 000W ZRn n n n n with <CR><LF>
s/(0\.000E\+00(\s{1,})){7}//; # Replace 7 x 0.000E+00 with nothing
.
.

Reply