Shell Programming and Scripting

Hello, there. I have a file that's a horrible, horrible mess. (Basically, it's an export from a firewall config.) The people who generated the file didn't think that putting a newline in the middle of a hostname would ever be a problem. It is.

Here's an example of the stuff in the file:

Code:

Entity name: Random_Subnet
02F7B97A-9DA1-4393-A7E3-7F367356190C
Network address
192.168.151.0
Random_Subnet
Netmask
255.255.255.0
Entity name: AccessApplications_Group
E486A584-E708-4DB3-809E-F5685F2215E5
Access internet applications
Network Entity
Random_Subnet,Random_Subnet2,Random_
Subnet3,Random_Subnet4,Random_Subnet
5,Random_Subnet6

The part with the network address I have parsed, all well and good. But the second "entity name" entry, where there's a group, you can see that it puts line breaks in random places. There's no rhyme or reason - there doesn't have to be a special character or anything, it just puts that there.

This is the string I used to get it to that point:

Code:

tr -d '\t' < VPhosts.htm | sed 's/<.*>//g' | sed '/^$/d' | egrep -v "false|Symantec Gateway Security|Description|Locked|Read only|Caption|Sequence Revision|MAC address|UUID"

There's further string manipulation I do afterwards to get it into a proper comma-separated value file. But right here is the step where I need help.

I want to take every line which has commas in it, remove the line break from the end, and concatenate it with the line after. I want this only to happen on lines with commas, everything else should remain intact. I've tried everything in my limited skill set with sed, awk, perl and tr of which I could think, and I can't get it right. The closest I came was:

Code:

for i in `cat test.out`; do if [[ $i == , ]]; then echo -n $i; else echo $i; fi; done > test2.out

But that also put in a newline wherever there was a space, so it didn't help, either.

Can someone give me a little one-liner that will do this? It sounds so simple, but it has gone way beyond my meager ability at scripting.

Add empty line between Entiry session.

Code:

$ awk '{print (/^Entity name:/)?"\n"$0:$0} ' urfile

Entity name: Random_Subnet
02F7B97A-9DA1-4393-A7E3-7F367356190C
Network address
192.168.151.0
Random_Subnet
Netmask
255.255.255.0

Entity name: AccessApplications_Group
E486A584-E708-4DB3-809E-F5685F2215E5
Access internet applications
Network Entity
Random_Subnet,Random_Subnet2,Random_
Subnet3,Random_S

For your problem, can you provide more examples?

This will find the lines with commas and create a single line as long as there are not more than three lines with commas in a row

Code:

sed '/,/N;s/\n//;/,/N;s/\n//' file