cannot loop through log file with repetitive transactional data using cygwin on windows 7
 

I have a log file with transactional data of about 1800 transactions delimited by "<<< doInit" to start each transaction and ">>> processRequest" to end each transaction with no spaces between transactions.I need to be able to separate and parse out specific values for each transaction. The script below I am using will only find the first transaction. How do i tell the script, I want to run a set of commands on all 1800 transactions individually.

$ cat readscript
#!/usr/bin/sh
set -x
echo "what is the input file"
read f


for value in $f
do
t=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 traceuid | awk -F ' ' '{print $5,$6}' | sed -n '2p' | cut -c 11-`
tx=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 transactionid | awk -F ' ' '{print $1,$2,$5,$6}' | sed -n 3p`
r=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 retailername | awk -F ' ' '{print $1,$2,$5,$6}' | sed -e '$!{h;d;}' -e x`
rc=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 responsecode | awk -F ' ' '{print $1,$2,$5,$6}' | sed -n '3p'`
time=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 datetime | sed -n '3p'`
tt=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 milli | sed -n '2p' | awk -F ' ' '{print$11}'`
ra=`cat $f | tr '<' '\n' | tr '>' '\n' | egrep -i -B1 path | sed -n '2p' | awk -F ' ' '{print$6}' | cut -c 6-`
echo 'TraceUID='$t, 'TransactionID='$tx, 'RetailName='$r, 'ResponseCode='$rc, 'Merchanttime='$time, 'totalmillisec='$tt, 'requestaction='$ra>> script1out.txt 2>&1
done

Welcome to LQ!

When posting code at LQ, please use [CODE][/CODE] tags around the code. It improves readability and maintains white space.

I think you need to study awk a little more. What you are doing with 'cat', 'tr', 'egrep', 'sed' and 'cut' can all be done within awk.
For example:
would select lines containg the string 'traceuid' and print just the fifth and sixth fields delimited by space characters. (You do not need "-F ' ' " option to awk as it is the default.)
Where you have used 'cut', look at 'substr' in awk.

If you want further help, posting a sample of the data, suitably redacted for privacy, would be of benefit.