Shell Programming and Scripting

Not sure what i was doing their Turk, that does work well, long day !!!

I've got the extra figures i was after, is there a simple way i can sum up these temp values at the end so i can see total success pass purchase for the day and total errors? What i mean is, an overall count for the success column, and a count for the other column combined (2 & 3)


PASS success failed already_booked internal_server_error

EPASS011 27 4 0 0
EPASS009 11 0 0 0
EPASS07 102 3 0 0
EPASS012 40 1 2 14
EPASS078 1545 18 0 3
EPASS013 627 6 0 0
EPASS013 1456 12 0 0
EPASS013 1266 4 17 4


Many thanks guys

Jeffers

I didn't saw this operator before in any book of perl but I had tried it but it does not work see below:-

"my perl version is v5.6.1"

Code:

perl -le '%x=qw(10 0 20 0); $k=20; printf("\nKey   = %5s\nValue = %5s\n\n", $k, $x{$k} // -9);'

o/p:-
Search pattern not terminated at -e line 1.

Code:

perl -le '%x=qw(10 0 20 0); $k=20; printf("\nKey   = %5s\nValue = %5s\n\n", $k, $x{$k} || -9);'

o/p
Key   =    20
Value =    -9

It's a new operator in version 5.10 and higher.

perl5100delta - search.cpan.org

tyler_durden

Hi Jeffers,

So to sum the values in the columns, you can modify the final print statement a bit like so, or sum it up in the main processing loop:

Code:

foreach (sort keys %resultsHash) {
    $temp = $resultsHash{$_};
    print "$_ : @$temp\n";
    $sum1 += @$temp[0];
    $sum2 += @$temp[1] + @$temp[2];
}
print "Success [$sum1]\nFail [$sum2]\n";

Sample run:

Code:

eturk-linux 09:41:10 (~)> ./filter.pl data.dat
EPASS013 : 1 0 4
EPASS088 : 2 1 0
EPASSV09 : 1 2 0
Success [4]
Fail [7]
eturk-linux 09:41:15 (~)>

As a general unsolicited piece of advice, I recommend strongly that if you will be using more Perl in the future, you read some documentation of the language. The real deal is here:

Learn Perl - www.perl.org

I also like the Perl monks site for more advanced topics:

http://www.perlmonks.org/?node=Tutorials

There are a ton of resources out there for the inquisitive Perl person.

Cheers,
turk451

Thanks Turk,

Spent the week reading the sites recommended, still find it confusing but have now massively improved the script to incorporate other searches, not sure if its the best practise but im getting somewhere, thanks for the helpful advise.

Jeffers

Hi Turk,

Wondered if you could offer some advise on arrays within arrays.

I would like to amend the script so that it doesnt count duplicate entries, what im seeing is a large error count when in fact if i run a sort and pipe it to uniq it turns out to be only a few users, is their a way i can modify the script to just get uniq users by the mobile number? Do i need to create another array within this array?

thanks

Hi Jeffers,

I would probably pre-filter the data using shell commands. One command you could use just to count the number of unique users would be:

Code:

cut -f 4 -d',' data.dat | sort | uniq | wc -l

if the data file just contains a lot of duplicated entries that you want to pre-filter (assuming they are exact duplicates), just skip the cut step:

Code:

sort data.dat | uniq > data.dat.new

Then you can just run the regular script on the cleaned-up data set.