[SOLVED] replace txt between 2 strings.

fur · 08-04-2010, 10:59 AM

I am trying to replace a section of a file between the first instances of the strings {}, with the contents of another file.

Example of the format of the file I'm trying to modify

Code:

Servername=something.com

     hosts {
       macaddress1
       macaddress2
       macaddress3
       macaddress4
       macaddress5
       macaddress6
     }
other stuff that
doesn't matter { }

So I would like to replace

Code:

     hosts {
       macaddress1
       macaddress2
       macaddress3
       macaddress4
       macaddress5
       macaddress6
     }

With data from another file that contains,

Code:

       macaddress7
       macaddress8
       macaddress9
       macaddress10
       macaddress11
       macaddress12

Tried to set,
IFS=""

Then captured all the "macaddress#"s to a variable and used sed to swap

sed "s/$CURRENTDATA/$NEWDATA/" filename

However I get 1 of 2 errors,

Using a small number of macs in "$NEWDATA"

sed: command garbled: s/ macaddresshere

Or when using a large number of macs in the $NEWDATA variable get

bash: /usr/bin/sed: Arg list too long.

Any other ideas on how to replace a large block of txt in one file with another large block from another?

Thanks

ghostdog74 · 08-04-2010, 11:09 AM

Code:

$ cat file1
       macaddress7
       macaddress8
       macaddress9
       macaddress10
       macaddress11
       macaddress12

$ cat file
Servername=something.com

     hosts {
       macaddress1
       macaddress2
       macaddress3
       macaddress4
       macaddress5
       macaddress6
     }
other stuff that
doesn't matter { }

$ awk -vRS="}" 'FNR==NR{s=s$0;next}/hosts/{ gsub(/{.*$/,"");$0=$0"{\n"s}RT{print}' ORS="}" file1 file
Servername=something.com

     hosts {
       macaddress7
       macaddress8
       macaddress9
       macaddress10
       macaddress11
       macaddress12
}
other stuff that
doesn't matter { }

fur · 08-04-2010, 03:26 PM

Hrm, I tried that on a Debian, and Solaris machine using awk and nawk, and I don't get any output.

Just drops to the next prompt.

Code:

$ cat file1
       macaddress7
       macaddress8
       macaddress9
       macaddress10
       macaddress11
       macaddress12
$ cat file
Servername=something.com

     hosts {
       macaddress1
       macaddress2
       macaddress3
       macaddress4
       macaddress5
       macaddress6
     }
other stuff that
doesn't matter { }

$ awk -vRS="}" 'FNR==NR{s=s$0;next}/hosts/{ gsub(/{.*$/,"");$0=$0"{\n"s}RT{print}' ORS="}" file1 file
$

fur · 08-05-2010, 09:38 AM

Using it with gawk works.

Thanks

fur · 08-28-2010, 01:31 PM

How could you use the same command, and have it match only the first instance in a file?

grail · 08-29-2010, 12:33 AM

Here is an alternative:

Code:

awk 'f{getline < "f2"; print}!f;/hosts/{f=1}/}/{f=0}' file

Quote:

How could you use the same command, and have it match only the first instance in a file?

First instance of what? Do you mean the first mac address? Or do you have more than one 'hosts{}' grouping?

crts · 08-29-2010, 02:08 AM

Hi,

this will replace anything inside the first pair of '{}', regardless if it starts with
hosts {
guests {
or whatever {

Code:

sed -e '/{/ {P;r file2
 :mark /}/ ! {N;b mark}; /}/ {s/.*\n\([         ]*}\)/\1/;};: endloop /\$/ ! {N;b endloop}; }' file

It is important that there mustn't be anything after 'r file2'. Otherwise it will be interpreted as part of the filename. Yes, it's a oneliner that spans over two lines.
The [] contains a tab and space.