awk help: summarize list when column X changes value

[Stuart Abramson]

This is the output from the EMC symdisk command:

DA-7B 07B D 1 SEAGATE CUDA180 24 173017 450 173145
DA-7B 07B D 2 SEAGATE CUDA180 20 173017 386 173145
DA-7B 07B D 3 SEAGATE CUDA180 20 173017 386 173145
DA-8B 08B C 5 SEAGATE CUDA180 20 173017 386 173145
DA-8B 08B C 8 SEAGATE CUDA180 20 173017 386 173145
DA-8B 08B C 9 SEAGATE CUDA180 20 173017 386 173145
DA-8B 08B D 0 SEAGATE CUDA180 21 173017 383 173145
DA-8B 08B D 1 SEAGATE CUDA180 25 173017 5946 173145
DA-8B 08B D 2 SEAGATE CUDA180 21 173017 383 173145
DA-8B 08B D 3 SEAGATE CUDA180 20 173017 386 173145
DA-9B 09B C 1 SEAGATE CUDA180 21 173017 383 173145
DA-9B 09B C 2 SEAGATE CUDA180 20 173017 386 173145
DA-9B 09B C 3 SEAGATE CUDA180 20 173017 386 173145
DA-9B 09B C 4 SEAGATE CUDA180 20 173017 386 173145
DA-9B 09B C 5 SEAGATE CUDA180 20 173017 386 173145
DA-9B 09B C 8 SEAGATE CUDA180 20 173017 386 173145
DA-9B 09B C 9 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B C 4 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B C 5 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B C 8 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B C 9 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B D 0 SEAGATE CUDA180 21 173017 383 173145
DA-10B 10B D 1 SEAGATE CUDA180 21 173017 383 173145
DA-10B 10B D 2 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B D 3 SEAGATE CUDA180 20 173017 386 173145
DA-10B 10B D 4 SEAGATE CUDA180 20 173017 386 173145

I want to summarize when the "DA" changes.

I want to say:
DA-07B has 4 x SEAGATE CUDA180 Disks.
DA-08b has 4 x .. etc.

How do i do that in "awk"? Is there some way to say "when $1 changes, count up.."? Or do I have to save the value and check every time, etc.?

[James R. Ferguson]

Hi Stuart:

Try this:

#!/usr/bin/awk -f
{if (FIRST!=1) {FIRST=1;LAST=$1}}
{if ($1!=LAST) {print LAST," count = ",CNT;CNT=0;LAST=$1}}
{print;CNT++}

Regards!

...JRF...

[Steve Steel]

Hi

Keep it simple

cat tmp/in|cut -f1 -d " "|sort -u|
while read line
do
let xx=$(grep $line tmp/in|wc -l)
echo $line has $xx $(grep $line tmp/in|
head -n 1|cut -f5-6 -d" ") disks
done


tmp/in is data file containing list shown


Steve Steel

[Rodney Hills]

For a simple count where the disk type doesn't matter-
cut -c1 yourfile | sort | uniq -c

Perl can provide a more complete routine-
perl -an -e '$tot{$F[0]}{"$F[4] $F[5]"}++;END {for $DA (sort keys %tot) { for $dsk (sort keys %{$tot{$DA}}) { print "$DA has $tot{$DA}{$dsk} $dsk Disks.\n";}}}' yourfile

HTH

-- Rod Hills

[Peter Nikitka · Accepted solution]

Hi,

just shorter and without awk:
cut -f1 -d " " /tmp/x | sort | uniq -c

mfG nik

[James R. Ferguson]

Hi Stuart:

BTW, you never evaluated the responses in your thread below and I would like to know if you found them helpful. Thanks.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=976274

Regards!

...JRF...

[Sandman!]

Stuart,

How about this awk construct:

awk '/^DA-[0-9]*B/{line=$5" "$6;if(NF==1){curr=$1;count[curr]++} else{if(curr==$1)count[curr]++;else{curr=$1;count[curr]++}}} END{for(i in count) print i " has", count[i] " x " line " Disks"}'

cheers!

[Sandman!]

Matter of fact the code is easier to understand if put into its own script:

=====================myawkscript=======================
/^DA-[0-9]*B/ {
line=$5" "$6
if ((NF==1)||(curr==$1)) {
curr=$1
count[curr]++
}
else {
curr=$1
count[curr]++
}
}
END{for(i in count) printf("%s has %d x %s Disks\n", i, count[i], line)}
=======================================================

# symdisk | awk -f myawkscript

cheers!