Re: Re (tilly) 2: Dynamic array names

[arturo]

You're right. I could have been more explicit in the original code, but I just wanted to give the gist, and so left large (and potentially important) hunks out. Of course I'd use my @array when I got it from "Thing" (which I'll assume is a line from a file which we're splitting to get an array) , à la :

my @list_of_arrays;
while (<FILE>) {
    my @array = split "\t", $_;
    push @list_of_arrays, \@array;
}

Because that gives you a (reference to a) fresh @array each time through the loop, instead of adding yet another reference to the same global array.

As tilly has pointed out, only giving *parts* of the answer may not be helpful ... So, as they said in that long-ago decade, let's be careful out there =)

Peevish answer to the interview question: "What, don't you people use strict around here? =)"

Philosophy can be made out of anything. Or less -- Jerry A. Fodor

[Fastolfe]

I would give even more bonus points if the candidate was able to work without the temporary @array variable at all:

my @list_of_arrays;

while (<FILE>) {
   push @list_of_arrays, [ split /\t/ ];
}

[tilly]

I would subtract considerable points for missing the fact that you were breaking the spec of what was asked for in at least 3 different ways! OTOH anyone who came up with this would impress me:

my @data = map [/[^,\n]*/g], <CSV>;

Now for shame, find at least 3 ways in which you broke the spec. (Having the wrong array name and the wrong filehandle don't count. I am looking for egregious data errors.)

[Fastolfe]

Had I been responding to you and not arturo (heh), I would have chomped the input data first, used a comma instead of a tab and a negative LIMIT argument to split to preserve trailing null fields. Did I miss anything else?

I might also have used /[^,\n]+/ in that map instead of a *, which introduces spurious empty fields by succeeding emptily upon encountering a delimiter.

[tilly]

I would give bonus marks for that peevish answer (but only if you can explain why the fix mattered). :-)

The point of this question is to see how the person debugs a simple problem. It is our responsibility to make sure that people who join learn the basic techniques we rely on to avoid them in the first place. (Including strict.)