UNIX for Dummies Questions & Answers

Dear all, will appreciate your help with this (supposedly simple) thing.
How to use sed to replace a part of string like this:

(TAGthenwhatever:

into (TAG:

meaning that '(TAG' (known, can be specified to sed and is marked with the '(' at the beginning) will stay and the 'thenwhatever' (until the first occurrence of ':') will go. The whatever can really be whatever but it is not interrupted with blanks (continuous string).

thanks a lot!

Can you please provide a better example, including original text and your desired output?

the original string would be

....1.08305,((THDEA1_1_PE708:0.47029,...

and, provided that I know the name THDEA,

the line would become 1.08305,((THDEA:0.47029,

The name is known every time I want to make the excision after it. But the surroundings will vary.

Try this:

Code:

sed 's/\(.*THDEA\).*\(:.*\)/\1\2/'

thanks- really unix is elegant

-----Post Update-----

Hey, here i'm again:
it seems in strings where there are more than one ':' after the name (in this case, AQAEO) this command eats up the line up to the last ':' in the string instead of excising the text between the name and the first occurrenc of ':' after it.

I attach the initial and the resulting files. I tried to play with the '.*' part of the command but the result's the same. U have an idea why?

Thanks in advance.

This command should remove everything between the given pattern and the first colon:

Code:

sed 's/\(.*THDEA\)[^:]*\(:.*\)/\1\2/'

Regards

Why don't you simply replace (TAGthenwhatever: with (TAG: ?