trying to isolate a block from my logs with awk, help please! -

i trying write awk script can block contains word "error", out of long log file.

basically log file contains actions performed, , when 1 of these fails, add under action error line, saying wrong.

i can isolate error line doing grep "error:", missing command given since printed before error line, , not know how many lines before, cannot arbitrarily "print 10 lines precede word "error:"

i've figured out sort of scheme thou; each of block contain error lines starts in same way ("processname"), followed command , other parameters, each 1 on different line, , last line empty line.

so idea use block awk, can "processname" string, start print lines 1 one until reach empty line, , pipe printed result trough grep see if there word "error:" in there; if there error redirect on file , append whole block, otherwise continue next block , same thing.

now if can hand task; since not know how achieve this; i've looked @ awk , seems right tool job (i can write shell script task), if think there better way in shell script, ears :)

thanks in advance help!

update: scripts; i've got 1 dennis work prints same block more once, if there more 1 error entry in each block; while example bellisarius not return line.

i've added example of log looks like, when introduce error (there empty line @ end, cannot add if put text in tag code):

processname id=1231 command:"ls -l a" hash "gkfsmgflsdmgklsdmfldsmfklmdsflkmsdflmsdflkmsdflkmsdfklsdmfklsdmfklmsdfklmsdklfmsdklmflksdmflkdsmfkldsmfkldmslfmdslkfmklsdmflksdmfklsdmfkldmslfkmslfkmsdlkfm" /filename/compileme.c:20: error: directory not exist /filename/compileme.c:20: error: incorrect parameter 

a regular block looks same, not have error: part obviously.

hope makes more clear, again!

you may try:

begin                {flag="no";k=0}  /processname/        {flag="no";k=0}  /processname/,/^$/   {a[k++]=$0;                       if(match($0,"error")!= 0) {flag="yes"};                      }  /^$/                 {if (flag=="yes") {flag="no";                                         ( i=0; i<k; i++ ){print a[i]}                                        print "-------";                                       };                       ( r in ){delete a[r]};k=0;                      } 

test:

input:

processname adasd asdasd  processname err  error   salutti  processname no err   aaa no err   processname   

output:

processname err  error   salutti  ------- 

running @ ideone here

edit

on comment having sometimes empty record before error record, can solve pre-processing log files following awk script deletes empty lines before error messages:

/^$/ {getline; if($0 !~ /error/) print ""}   {print}   

and running main script on output of one.