How to sort fields of unequal lenght

[Frank de Vries]

Hi,
I have a file with source and destination ip@ and the traffic in bytes, delimited by ;
10.214.5.205;10.218.25.220;3568
10.214.4.235;10.217.25.26;250
10.214.3.2;10.216.25.26;1002506
10.214.3.20;10.219.25.200;4569870125
10.214.2.215;10.215.25.26;4024
10.214.1.205;10.214.25.26;10250

I want to sort on the traffic field to
see the top usages

I tried
cat test.sh | awk -F ";" {' print $3";"$2$1 '} | sort -k1,1
but that gives unsatisfactory output:
1002506;10.216.25.2610.214.3.2
10250;10.214.25.2610.214.1.205
250;10.217.25.2610.214.4.235
3568;10.218.25.22010.214.5.205
4024;10.215.25.2610.214.2.215
4569870125;10.219.25.20010.214.3.20

What we need to see is this:
4569870125;10.219.25.20010.214.3.20
1002506;10.216.25.2610.214.3.2
10250;10.214.25.2610.214.1.205
4024;10.215.25.2610.214.2.215
3568;10.218.25.22010.214.5.205
250;10.217.25.2610.214.4.235

I tried in perl with sort() but
it can't handle the field mappings.

So know I have run out of ideas
Thanks for your help

[James R. Ferguson]

Hi Frank:

Add the '-n' switch to your sort:

# awk -F ";" {' print $3";"$2$1 '} /tmp/data |sort -kn1,1

Regards!

...JRF...

[Peter Nikitka]

Hi,

I think it is best to add the field delimiter ';' and the numreic option:

awk -F';' '{print $3";"$2$1}' test.sh | sort -t';' -nrk1,1

mfG Peter

[James R. Ferguson]

Hi (again) Frank:

Yes, I agree with Peter, and I missed that. Add the field delimiter to the 'sort' too.

Regards!

...JRF...

[Frank de Vries]

Thanks for jogging my memory !!
Solution was so close and yet so far.

Anyone know how to do this in perl examples ??

[H.Merijn Brand (procura · Accepted solution]

Trie to find the example of how to write Schwartian Transsforms

my @sorted =
map { $_->[0] }
sort { $b->[9] <=> $a->[9] ||
$a->[1] <=> $b->[1] }
map { [ $_, split m/\D+/, $_ ] }
@unsorted;

Extend the sort to your content

Enjoy, Ha ve FUN! H.Merijn