Hello i have a file that contains date and time format,
what i want to do is to replace the seconds field with zeros(00)

Example:

Before:
After:
I start with a sed command but i replace, minutes and seconds to 00

sed -e 's/:[0-5][0-9]^:/:00/g' file.txt

I have read several pages but i doesn't understand how to do this


Can anybody help me with this problem ??

the command works as:
s/<search regexp>/<replacement>/
so you need to search for :digitdigitspace and replace with :00 (including space at the end)
So instead of ^: at the end of your search expression:
s/:[0-5][0-9] /:00 / would be better (g is not required)

Try:

Please note the spaces after 9] and 00

Even simpler:

Hello Thanks for really fast answer :-)

When i try this commands i get the same results :-(




andreas@server:~$ sed 's/:[0-5][0-9] /:00 /' file.txt
2011-01-27 03:50:13 156109680 226945640
2011-01-27 03:51:02 156109960 226945960
2011-01-27 03:52:01 156110250 226946280
2011-01-27 03:53:45 156110540 226946600

andreas@server:~$ sed 's/:.. /:00 /1' file.txt
2011-01-27 03:50:13 156109680 226945640
2011-01-27 03:51:02 156109960 226945960
2011-01-27 03:52:01 156110250 226946280
2011-01-27 03:53:45 156110540 226946600


/Andreas H

Can you post a sample of the file wrapped in the code tags?
Are those spaces or tabs?

What version of sed are you using?

you can use ^: - between [ and ] instead of space:
sed 's/:..[^:]/:00 /'
maybe it works

Thanks everyone this solution work perfect.

/Andreas H

If someone helps you, or you approve of what's posted, click the "Add to Reputation" button, on the left of the post.
If you want to highlight your sentences just use bold or italic text, but not code.
I'm glad to help you.

I think I would use awk instead. It just seems cleaner, as you can target the time field only:


And here are some references for you.

sed:
http://www.grymoire.com/Unix/Sed.html
http://sed.sourceforge.net/grabbag/
http://sed.sourceforge.net/sedfaq.html
http://sed.sourceforge.net/sed1line.txt

awk:
http://www.grymoire.com/Unix/Awk.html
http://www.gnu.org/software/gawk/man...ode/index.html
http://www.pement.org/awk/awk1line.txt
http://www.catonmat.net/blog/awk-one...ined-part-one/

nice, and if you want to write less:
is enough

sycamorex, you hit the nail in the head. I've added CODE tags and they reveal the presence of TABs. Andreas H and pan64 you're kindly invited to use CODE tags in the future, for both the questions and the answers. Thank you.

Of course. But my general policy is to give the traditional explicit syntax for new scripters. Posting advanced shortcuts, at least without including detailed explanations of them, tends to confuse more than help. They also tend to be rather situation-specific and not easily adaptable to more generalized use.

More experienced users can usually figure out how to shorten it on their own.


For the record, the above code works because of the basic awk 'pattern { commands }' syntax. If you leave the bracketed commands off and only have a pattern, then it defaults to printing the line on a true evaluation. In this case, the sub function returns true when a valid substitution is made, and so the line prints.

Note though that because this relies on the status of sub to perform the action, if there happened to be any lines in the file that didn't contain a "$2" field, they wouldn't get printed. But the explicit declaration of the action in my version prints every line regardless of the substitution result.