Reading input record from inside nawk

Hi friends,

I have small query with reg to awk search pattern..

below is my sample file and code which i tried..

$ cat file.txt
xxx,yyyyy,messageID,sha
xxxx,errorcode,messageID,name

in the above sample file - let assume I know the errorcode(2nd record) using which I want to find out sha record(1st line last column), the key is messageID which would be common for both the record.

I have tried following awk syntax but failing

nawk -F"," '{if($0~/errorcode/) {var=$3;if($0~/var/ && $0!~/errorcode/) print $NF}}' file.txt

Thanks
SHa

---------- Post updated at 03:51 AM ---------- Previous update was at 03:47 AM ----------

Just to add above statement for better understanding...the record which I am searching in above syntax can be anywhere not always one line above or below...thats y i have not tried getline etc....

Thanks
Sha

something like:

#  nawk -F"," '{($2~"errorcode")  ? t[$3]=$2 :  s[$3]=$4 }END{for (x in t){print x,t[x],s[x]}}' infile
messageID errorcode sha

should be close to what you're after.

HTH

Cool...worked perfectly...would you please also explain bit on for loop part how does it work....which is bit confusing me.........

Anyway...Thanks a lot..

Thanks
SHa