Shell Programming and Scripting

@RRVARMA
Your script give me
real 0m2.997s
user 0m0.980s
sys 0m1.624s

Replace

Code:

	field7=`echo $j|cut -d "|" -f7`
	field13=`echo $j|cut -d "|" -f13`
	field14=`echo $j|cut -d "|" -f14`
        if [ "${i}" == "${field7}|${field13}|${field14}" ]

by:

Code:

 if [ "${i}" == "$(echo $j | awk 'BEGIN{FS=OFS="|"}{print $7, $13, $14}')" ]

and you will save %43 procesing time using the same resources.
real 0m1.683s
user 0m0.612s
sys 0m0.888s


@cooldude solution demands more memory but is faster.
real 0m0.211s
user 0m0.036s
sys 0m0.004s


@Franklin52 awk solution is the best, but uses a significant amount of memory
real 0m0.005s
user 0m0.000s
sys 0m0.004s

If you really want to speed things up, store the contents of the first file in an array. Then read the records of the second file one by one and compare the seventh, 13th, and 14th records against the array. You can perform the binary search on the array since the array is already sorted. The problem with fpmurphy's solution was that he was constantly invoking cut and reading the file again and again. In my solution, I used ksh's functionality to split the fields in the file, but I am still reading the file again and again. I'm betting the solution I'm proposing will be even faster than Franklin52's.

you're lucky you have not had your thread closed for cross posting if this is he third thread you have started for the same topic

Since this thread seems to be the longest of the three that he had started, I am keeping this open and closing the others.
RRVARMA, please adhere to the rules. No double posting or bumping up the posts.
Here are the other threads.
https://www.unix.com/shell-programmin...wk-script.html
https://www.unix.com/shell-programmin...ther-file.html

Hi,

Thanks cooldude, but your script is fetching not only the unique records but also the other records as well.. I'm confused..

But its quite faster than the other scripts which i tried.

Thanks vino, for closing other two threads.. Sorry that i didn't know how to close those threads.

I recommend Franklin's solution. But here's another version:

Code:

#!/bin/ksh

if [ $# != 2 ] 
then
exit 1
fi

while read i
do
    while read j
    do
	set -A array $i
	field7=${array[6]}
	field13=${array[12]}
	field14=${array[13]}
        if [ "${i}" == "${field7}|${field13}|${field14}" ]
	then
	print "$j"
	break 
	fi
    done < $2
done < $1

It is the same as the script you posted in your first post, except that I am not using cut to extract fields.