Re: Optimizing a sort function (wrap-around alpha)

I'd write that as:

my @sorted = ((sort grep {$_ ge $top} @unsort),
              (sort grep {$_ lt $top} @unsort));
[download]

Or, if you really insist on a single sort:

my @sorted = map {substr $_, 1}
             sort 
             map {sprintf "%s%s", (0, 1, 2)[$top cmp $_], $_}
             @unsorted;
[download]

Neither technique uses a sort block, and that usually means it's more efficient than a method using a sort block.

Comment on Re: Optimizing a sort function (wrap-around alpha) Select or Download Code

Replies are listed 'Best First'.
Re^2: Optimizing a sort function (wrap-around alpha) by Anonymous Monk on Apr 15, 2005 at 11:33 UTC
And yet another way: `my %h; push @{$h{$_ cmp $top}}, $_ for @unsorted; my @sorted = (@{$h{0}}, sort(@{$h{1}}), sort(@{$h{-1}}));` [download] It's using a single pass bucket-sort as first step. Again note the absense of a sorting block.	[reply] [d/l]
Re^2: Optimizing a sort function (wrap-around alpha) by bugbugbear (Initiate) on Apr 15, 2005 at 14:27 UTC
For the latter solution, you must have meant `[$_ cmp $top]` instead of `[$top cmp $_]`.	[reply] [d/l] [select]