jumping from one line to another

UNIX Solaris
1

Hi,

Thanks

2

Hi Rocky1954,
Are you looking to find a unique list of hostnames from the file?
if so use the below code.
Assume the file containing that output is called sample.txt

awk -F"=" '{print $7}' sample.txt|awk -F")" '{print $1}'|sort -n|uniq -c|awk '{print $2}'

Hope this helps
Vj

3

Thanks for your reply

4

Assuming multiple lines, that might not all have host data, and making it a bit simpler:

awk '
        /HOST=/ {                # for each input line with host name
                sub( "^.*HOST=<", "" );  # delete everything up to host name
                sub( ">.*", "" );        #delete everything past hostname
                print;
        }
' <input-file | sort -u

If you need the hostname contained within < and > in the output, remove those characters from the substitution pattern.

mvijayv's example was basically doing the same thing, but using more processes than necessary. First awk set the field separator to = and printed the 7th field ($). Second awk set the field separater to close paren and printed the first field. Then the output was passed through sort and unique to sort the list and remove duplicates.

5

This will work if the host is proceeded by "HOST=" on each line, it only uses 1 awk process.
It also supports multiple HOST= on 1 line

awk -F"[=,(,)]" '
{
  for (i=1;i<=NF;i++) {
    if ($i == "HOST") H[$(i+1)]++;
  }
}
END { for(host in H) print host; }' logfile

It uses awk with "=" "(" and ")" as delimiters and uses the field following HOST as as has array index. After the whole file is read in the contents of the hash array is printed.

6 
awk -F"[()=]" '{print$16}' infile | sort | uniq
7

Thanks

8 
awk -F"[=() ]" '
{
  if(NR==1) cur=$1
  if (cur != $1)
  {
      for(host in H) print cur,host,H[host];
      delete H;
      cur=$1;
  }
  for (i=1;i<=NF;i++) if ($i == "HOST") H[$(i+1)]++;
}
END { for(host in H) print cur,host,H[host] }' logfile
9 
awk -F'[(=)]' '{for(i=0;++i<=NF;){if($i=="HOST"){_[$1 OFS $(i+1)]++}}}END{for(i in _)print i,_}' file
10 
sed 's|(.*HOST=\([^)]*\).*|\1|' input

[ctsgnb@shell ~]$ echo '24-APR-2009 (CONNECT_DATA=(SERVER=DEDICATED)(SERVICE_NAME=<service_name>)(CID=(PROGRAM=root)(HOST=<host_name>)(U SER=rock)))' | sed 's|(.*HOST=\([^)]*\).*|\1|'
24-APR-2009 <host_name>
[ctsgnb@shell ~]$
11

danmero, could put ' ' in -F list to avoid double space in output, result is not in date order - might need to sorting it after.
ctsgnb, no count, if more that 1 "HOST=" on line only first is output.