using sed to remove a strange token

jacksmash · 09-11-2008, 09:41 AM

How would one use sed to go about removing a token like: &word;

I've tried:

1. sed 's/&word;/ /'
2. sed 's/\&word\;/ /'

things like that.

Any suggestions?

CRC123 · 09-11-2008, 09:45 AM

Umm, they both worked for me exactly as you have typed them

jacksmash · 09-11-2008, 09:49 AM

weird...

do you think it may be a problem if they are not separated by a space? For example, what if you have asghasgpn&word;asdgnapsd as your input. Will it still remove it?

Thanks.

CRC123 · 09-11-2008, 09:52 AM

yep, sure does

jacksmash · 09-11-2008, 09:54 AM

good grief this is weird

CRC123 · 09-11-2008, 10:04 AM

certainly. Do you have any aliases set up for the sed command?

Code:

alias | grep sed

and what distro you running and what version of sed?

EDIT: nevermind version, I see it's probably ubuntu from your profile

jacksmash · 09-11-2008, 10:15 AM

actually, I just ssh'd to my school's server and I'm running it from there... so I'm not sure of the details. When I run the alias command as suggested, I get nothing.

jacksmash · 09-11-2008, 10:19 AM

oh.. and I'm not sure what version of sed either. I would use another tool like 'tr', but I don't think you can remove specific words with 'tr'.

CRC123 · 09-11-2008, 10:20 AM

Well if it's a school server, ask the system administrator what the heck is going on

. Have you tried the commands on your home box? Do you get different results?