sed remove statement

Shell Programming and Scripting
1

I am having some problems with sed, that I am hoping that I can get some assistance with. I am trying to remove two subsets of a string, and cannot figure out how to have it work.

Here is an example string:

[Wed Sep 24 08:47:46 2008] [warn] [client xxx.xxx.xxx.xxx] [28580] auth_ldap authenticate: user joe authentication failed; URI /svn/ [User not found][No such object]

I want to use sed to remove [Wed Sep 24 08:47:46 2008] [warn] including the [ and ]. I only want the first two of these removed.

Suggestions or pointers on doing something like this?

I cannot do this in the apache conf file, as this sed statement is actually so that I can send SNMP trap data, I do not want to change what is going into the log files.

2

including the [ and ]

I do not see this in your sample one line.
By the way, it would be easier to visualize, code and test if we could see a few more sample data lines.

3

Sorry, I meant the characters "[" "]".

A couple more line of the error log can easily be provided.. Here ya go..

[Wed Sep 24 08:32:58 2008] [warn] module authz_svn_module is already loaded, skipping [already loaded]
[Wed Sep 24 08:33:00 2008] [warn] module dav_svn_module is already loaded, skipping [already loaded]
[Wed Sep 24 08:33:00 2008] [warn] module authz_svn_module is already loaded, skipping [already loaded]

I just want the first two statements removed (date string, and error levels).

4

With awk you can remove the two blocks within the square brackets as follow:

 awk 'BEGIN{FS=OFS="]"}{$1=$2="";sub(/^\]* /,"")}{print}' file > newfile

Regards

5

Thanks!

I begged some of the other guys I know to assist me, and they gave me two other options as well.. Here they are for anyone else who needs something like this..

sed -e 's/^\[[^]]\] \[[^]]\] //'

perl -pe 's/^\[.?\] \[.?\] //'