Form a output by combining multiple lines

Question

I have a log file as below:

/export/home/got/logs/o2877612.01:job_1432787863184_159408 Message: org.apache.pig.backend.executionengine.ExecException: ERROR 2118: Input path does not exist: /user/got/distro /export/home/got/logs/o2877612.01:Module: SUMMARY_PIG 

I want to form a final log file by combining two lines as below. Above both lines start with /export/home/got/logs/o2877612.01:

/export/home/got/logs/o2877612.01:Module: SUMMARY_PIG :job_1432787863184_159408 Message: org.apache.pig.backend.executionengine.ExecException: ERROR 2118: Input path does not exist: /user/got/distro 

Could you help in the unix command for this?

Answer 1

With awk:

awk -F: '{if($1!=a){printf "\n%s", $0}else{$1=":";print}} {a=$1}' file 

Where:

• -F: uses : as delimiter.
• if($1!=a){...} If the a variable contains the same as the first field, print everything without a newline at the end,
• else{...}: else print all fields except the first one.
• {a=$1} Set the a variable used in the if clause.

Output:

/export/home/got/logs/o2877612.01:job_1432787863184_159408 Message: org.apache.pig.backend.executionengine.ExecException: ERROR 2118: Input path does not exist: /user/got/distro: Module SUMMARY_PIG 

I think you want line in the order as they appear in the logfile, not reversed as in your desired output.