pattern matching lines using the date, and then joining the lines

Shell Programming and Scripting
1

Hi Guys,

Was trying to attempt the below using awk and sed, have no luck so far, so any help would be appreciated.

Current Text File: The first line has got an "\n", and the second line has got spaces/tabs then the word and "\n"

   TIME           SERVER/CLIENT                      TEXT
02/03/2010 23:44:13 hostname02 hostname01  NBU status: 96, EMM status: No media is
                                          available

 01/03/2010 03:48:14 unknown hostname01  backup of  hostname01 exited with
                                          status 96 (unable to allocate new media for backup,
                                         storage unit has none available)

 22/02/2010 19:44:09 hostname02 hostname01  NBU status: 96, EMM status: No media is
                                          available

OUTPUT FILE:

TIME            SERVER/CLIENT                      TEXT
02/03/2010 23:44:13 hostname02 hostname01  NBU status: 96, EMM status: No media is available
01/03/2010 03:48:14 unknown hostname01  backup of client hostname01 exited with status 96 (unable to allocate new media for backup, storage unit has none available)
22/02/2010 19:44:09 hostname02 hostname01  NBU status: 96, EMM status: No media is available
2 
sed -i '$!N;s/\n[ |       ]*//g' file

Try the above code. Now the file will contain the text in the required format

3

You use the following command to achieve your requirement.

sed -ri '/^\n$/d;s/[ ]{2,}/ /g;s/[\t]{1,}/ /g;' <inputfilename>

-i is used to affect the changes to an original file.

If you don't want,then remove it and the output will be printed on stdout.

4

Thanks guys, but that didnt work, I am using the aix sed version, and the -i and -r options are not available, I tried both above commands with out the -i and -r options. no luck.

5

If -i and -r options aren't working.Then,you simply escape the curly braces in my command as follows.This will work.

sed '/^$/d;/^[ \t]\{1,\}$/d;s/^[ \t]\{1,\}\(.*\)/\1/g;s/[ ]\{2,\}/ /g;s/[\t]\{1,\}/ /g;' <inputfilename>

If -r option works,there is no need to escape the curly braces.

6

With awk...

awk '{gsub("^ *"," ",$0);printf ((/^ [0-9]/?"\n"$0:$0))}' infile
7

Thks guys for the response, really appreciate it. The sed command didnt work, but the awk did the trick. Thnks once again.