sed: Function cannot be parsed.
 

I have a shell script 'Test.sh'

echo "Enter The File Name"
read FILE_NAME

echo "Enter The First Column"
read Col

RowNum=2
Diff=4
Data=" "
ColNumFrom=`grep -i $Col Data.csv | cut -d "," -f2`
ColNumFrom=`expr $ColNumFrom - 1`

sed -i "${RowNum}s@\(.\{$ColNumFrom\}\).\{$Diff\}@\1${Data}@" $FILE_NAME >> NewFile


Data.csv is

AA01,
AA02,73
AB01,303
AB02,83


When I run the File for Col 'AA02', it works fine but for Col as 'AB01', is throws following error

sed: Function 2s@\(.\{302\}\).\{4\}@\1 @ cannot be parsed.

Please Help me solve the error...

what is the maximum line length sed can handle?

According to http://sed.sourceforge.net/sedfaq6.html

P.S. >> I'm curious why you are using the -i operator, which specifies an in-place edit, but at the same time you are directing output to another file?

For AAO2 nothing happens (well, not true, and empty NewFile gets created).

Plus you have the file-name hard-coded in the line with the grep ;}

Some versions of sed do not work with n greater than 255 in \{n\}
This regular expressions tutorial (written for Solaris) mentions the limitation.
http://www.grymoire.com/Unix/Regular.html#uh-8
Also
http://www.gnu.org/software/sed/manu...ar-Expressions

can i increase the length from 255 to 755 or something?

You could try building 755 up in sections using bash arithmetic. I don't know if this will work with all versions of sed. I only have GNU sed, which has an upper limit of 32767 for \{n\}.

is there no other way by which a sed command can edit the file beyond 255th character?

I thought that a previous post stated that there was no such limit?? What version of SED are you using?

Note that one post suggests the limitation only as part of the {n} construct--which I assumed to refer to a regex that is repeated n times.

Whatever the limitations actually ARE, you are not going to change them except by:
a. Using a different version
b. Getting the source code and compiling your own special flavor.

you are using the wrong tool for the job. You have a csv file where each column is separated by a delimiter(ie comma), its a whole lot easier to split the fields on commas, using tools like awk. (or the shell.)
eg

in man pages, there is a WARNING

The hold space is limited to 8192 characters.

What does this mean?

How many characters does the pattern in sed matches.. specifically \(.*\)?

So if this hold space is 8192, & i am trying to edit 2nd row, 302th character (where record/row length = 735 character, so i am editing 1037th character of the file), sed should not through any error.

But i am getting an error that function can't be parsed.

Hi,

the limitations of 255 mentioned earlier refer to the multiplier {} and not to the total buffer-space. Which version of sed are you using?

i am using sed.
I dont have any other version.:( (ssed GNU sed etc)

I can see the trailer as "HP-UX Release 11i: November 2000" in man pages, bt no version details.