Question

Dear all, I am writing a python program that is retrieving edifact log messages from a .gz file… An example of 2 logs are the following:

2009/03/02 12:13:59.642396 siamp102 mux1-30706 Trace name: MSG Message sent [con=251575 (APEOBEinMux1), len=2106, CorrID=000182C42DE0ED] UNB+IATB:1+1ASRPFA+1A0APE+090302:1213+0095JQOL2  2009/03/02 12:14:00.029496 siamp102 mux1-30706 Trace name: MSG Message sent [con=737 (APIV2_1), len=22370, CorrID=000182C42DE0ED] UNB+IATB:1+1ASIFQLFS+1ARIOFS+090302:1214+0122V11ON9 

I would like to write a regular expression able to match some field from the first line, some from the second and some other from the third …

Is there any way to write a regular expression to be used with GREP that matches field from consecutive lines ??

Thanks in advance !!!

Answer 1

With grep alone, I think this is not possible. I would suggest awk or perl in order to be able to save some context from previous lines.

In perl this gives something like:

#!/usr/bin/env perl  $isInLogSection = 'NO'; while (<>) {     if ( /siamp102/ ) {         # Start of log section: retrieve its ID         $isInLogSection = 'YES';         split;         $logSectionID = $_[0];     }      if ($isInLogSection eq YES && /len=/) {         # Retrieve value of len         ...     }      if ( /^$/ ) {         # End of log section         $isInLogSection = 'NO';     } } 

In awk this gives something like:

BEGIN { isInLogSection = 'NO'; } /siamp102/ { isInLogSection = 'YES'; logSectionID = $1; } /len=/ { if (isInLogSection == 'YES') { #retrieve len value } } /^$/ { isInLogSection = 'NO' } 

I am not 100% certain of the exact syntax. This is mainly a canvas for illustrating the principles.