[JOB] Perl Software Developer and Database programmer

[Peter Hickman]

Roger Burton West wrote:
>
> Given this problem, I'd probably ask "do you really want to treat
> duplicates in this way" - because the version that other people have
> come up with is much more commonly the one that's actually wanted...
>
> R
>
>   
What was wanted was an optimised version of the original code, that 
means it must produce the same results as the original code. "Hey I've 
optimised the payroll. If you put in more than one expense claim in a 
given period it only pays one of them and ignores the rest" is this "the 
one that's actually wanted".

I think not.

Given that we were not given any such information about the type of data 
that is held in alphas and betas then we cannot ignore the possibility 
of duplicates. We may also be in the position that the order of the 
results is also important so the sorted list with a binary search would 
probably fail for both duplicates and results ordering.