Shell Programming and Scripting

Hi Scrut,

Great ..

I need to understand your logic / code. If you have time can you share?

Code:

awk '/^... ... .. ..:..:/{if(p~s)print p; p=$0; next}{p=p ORS $0} END{if(p~s)print p}' s="ORA-07445" infile

Code:

awk '/^... ... .. ..:..:/{printf RS}{print NR ": " $0}' infile | awk '/ORA-07445/' RS=

Thanks

Hi, both methods attempt to create records that consist of a single log entries. Log entries start with a pattern of a date (3 characters, a space, 3 characters, a space, 2 characters, etcetera) such a pattern could be specified more precisely, but it is good to only use the precision that is required, and this will probably do...

Then once we have the start of the record we need to determine the end.

In the first method the end is signified by the start of the new log entry or when the end of the file is reached and the log entry is contained in the variable "p". If it matches "s" then it should be printed.

In the second method an empty line is created above the start of the log entry, so that the second awk can filter these entries by using the special case value of an empty record separator, which splits records if two consecutive newline characters are found...

I hope this helps,

S.

Hi Scrut,
Thanks a lot

Code:

awk '/^... ... .. ..:..:/
             {
     line1            if(p~s)
     line2            print p; 
     line3            p=$0; 
     line4            next
              }
                 
               {
                    p=p ORS $0
                
                } 
          END
             {
                if(p~s)  print p}' s="ORA-07445" infile

Please confirm if line 1-4 is if block. This means when if condition fails then control go to

Code:

 {
                    p=p ORS $0
                
                }

That's right.

Many thanks....