Find several different words and replace with one using sed.

Techno Guy · 07-06-2009, 04:57 AM

So as my title says, im trying to match multiple words and replace them all with one.

I have tried "test|just|this" in my regex checker app and it matches all 3 words, but when I try use it in the sed command it doesn't match any...
What am I doing wrong here?

Code:

echo This is just quick little test | sed 's/test|just|this/./i'

I know I could just make several sed commands for each word but that seems messy and not conventional.

I would appreciate some help here.

syg00 · 07-06-2009, 05:12 AM

Wrong tool for the job - spanners make poor screwdrivers.

Techno Guy · 07-06-2009, 05:13 AM

Quote:

Originally Posted by syg00

Wrong tool for the job - spanners make poor screwdrivers.

Ok, what would you suggest I use?

ghostdog74 · 07-06-2009, 05:15 AM

use awk

Code:

echo This is just quick little test | awk '{gsub(/test|just|this/,"")}1'

Techno Guy · 07-06-2009, 05:19 AM

Fantastic, thanks ghostdog74.
That does exactly what I need it to do.

I did have a quick look at the awk man pages (at the examples) but it didn't really seem like the right thing at the time, but now I know it is.

ghostdog74 · 07-06-2009, 05:30 AM

Quote:

Originally Posted by Techno Guy

Fantastic, thanks ghostdog74.
That does exactly what I need it to do.

I did have a quick look at the awk man pages (at the examples) but it didn't really seem like the right thing at the time, but now I know it is.

instead of the man pages, take a look at my sig for Gawk link.

syg00 · 07-06-2009, 05:31 AM

Another happy customer - cheers ghostdog74.
Dammit, you're going to make everybody awk users ....

Techno Guy · 07-06-2009, 06:07 AM

I'v opened tabs on nearly all of the links in your sig

Just one more quick question,
How would I do a case insensitive match? (with sad I could just go /i but it's not so easy with awk as far as i'v seen)

I had a little read from the link in your sig...

I don't get how it work? or how im supposed to use it.
....

Code:

echo This is just quick little test | awk 'tolower($1) ~ {gsub(/test|just|this|\ /,".")}1'

???

ghostdog74 · 07-06-2009, 06:12 AM

like i said , read the link. If you do, you will come across IGNORECASE. Set it to 1.

PMP · 07-06-2009, 06:13 AM

Well Sed also works take a look at this snippet

Code:

echo This is just quick little test | sed -e 's/test\|just\|this/./ig'


Output :
. is . quick little .

ghostdog74 · 07-06-2009, 06:25 AM

Quote:

Originally Posted by syg00

Dammit, you're going to make everybody awk users ....

one of the things good about learning awk, is that you can totally skip learning sed, since their capabilities overlap (and awk is a small programming language itself)

Techno Guy · 07-06-2009, 06:51 AM

Quote:

Originally Posted by ghostdog74

like i said , read the link. If you do, you will come across IGNORECASE. Set it to 1.

Yeah I did, but they also said

Quote:

"In general, you cannot use IGNORECASE to make certain rules case-insensitive and other rules case-sensitive, because there is no straightforward way to set IGNORECASE just for the pattern of a particular rule."

And since I may use awk as a case sensitive find&replace latter on, I would like to avoid changing things like that.

So, would you or someone else have any pointers as to how I would get it to work the other way (as stated in my previous post.)

PMP · 07-06-2009, 06:57 AM

Inspired by ghostdog :-)

Code:

echo This is just quick little test | awk 'IGNORECASE = 1 {gsub(/test|just|this/,".")}1'

syg00 · 07-06-2009, 06:57 AM

Quote:

Originally Posted by ghostdog74

one of the things good about learning awk, is that you can totally skip learning sed,

Funnily enough, I basically skipped awk and went straight from sed to perl.
Each to their own ...

ghostdog74 · 07-06-2009, 07:05 AM

Quote:

Originally Posted by Techno Guy

And since I may use awk as a case sensitive find&replace latter on, I would like to avoid changing things like that.

then you simply toggle IGNORECASE before you search for the next pattern.