Shell Programming and Scripting

I have a small query. I have a file containing the following lines

abcd<12></12>fdfgdf<12>sdfgfg<12>
sdfsdf<12></12>ytunfg<12>
hggfhf<12>rtysb<12>zdfgdfg<12>

Now I wish to delete ONLY the last occurance of string <12> from every lines of code. That mease my final output will be like this:

abcd<12></12>fdfgdf<12>sdfgfg<12>
sdfsdf<12></12>ytunfg
hggfhf<12>rtysb<12>zdfgdfg

The following sed command will remove the final occurrence of <12> at the end of each line - providing that <12> are the very last characters (i.e. no whitespace)

Code:

sed 's/\(^.*\)<12>$/\1/' oldfile > newfile

If there is whitespace after the <12> you could do something like

Code:

sed 's/\(^.*\)<12>[	]*$/\1/' oldfile > newfile

(Where a space followed by a tab are between the [ square brackets ] ).

Cheers
ZB

Sorry buddy but this command does not works for me.

Well it works for me using the sample data you provided, under three different flavours of Unix.

Why doesn't it work?

Are there any error messages?

Does the output not look the way you want?

Cheers
ZB

This command command neither gave any erros nor it give me any output, actually the file remains unchanged with this command. In fact the following command works for me

sed 's/\<12\>$//g' <filename>

dkhanna01 !

U mentioned yr required o/p as wrong in the initiial posting.
Please check ....

ZB postings work perfectly to replace the last instance of <12> in the file.

On some old versions of sed (I'm testing this with UNICOS 9.0 here...), you'll have to modify the sed thusly....

Code:

sed 's/^\(.*\)\<12\>$/\1/' oldfile > newfile

On newer sed (tests here: HP-UX 10.20, Linux circa 2.4, cygwin) the original examples work fine.

Maybe next time post the output of uname -a with your O/P?

Cheers
ZB