Help with sed regex to match words via a pattern.

logar0 · 10-24-2010, 11:44 AM

Hello everyone!

I have some lines of text and I want to detect words that meet some criteria with sed.

E.g. The line "tetris cat dog test cactus stereotype"

I want to detect the word that contains two (2) ore more t letters. That mean the words tetris, test and stereotype.

The following regex doesn't do the job because it can;t distinguish words very well.

/.*t.*t.*//

I thinks I have ti incorporate something like [^ ] in the regex but I wasn't successful...

Any help?

codegenie · 10-24-2010, 11:59 AM

I am not sure but why dont you try this::
a [a-zA-z0-9]
%%
{a}*[t]{a}*[t]{a}* {printf("valid ");}
%%

logar0 · 10-24-2010, 03:14 PM

No, unfortunately it doesn't work. Anyone else?

logar0 · 10-24-2010, 04:33 PM

Ok, I found the regex.

[ ]*[^ ]*t[^ ]*t[^ ]*