Inverse slices

Anonymous Monk has asked for the wisdom of the Perl Monks concerning the following question:

Replies are listed 'Best First'.
Re: Inverse slices by tybalt89 (Monsignor) on Nov 01, 2024 at 03:19 UTC
Here's a few ways (TIMTOWTDI). #!/usr/bin/perl use strict; # https://perlmonks.org/?node_id=11162546 use warnings; $SIG{__WARN__} = sub { die @_ }; # Is there a better way to get an inverse slice than to do something l +ike: my @array = ('aa'..'zz'); my @slice_idx = (6,13,42,66,69); # Slice my @slice = @array[@slice_idx]; use Data::Dump 'dd'; dd { 'array' => \@array, 'slice_idx' => \@slice_idx, 'slice' => \@slice }; my %slice_idx = map { $_ => 1 } @slice_idx; my @invslice_idx = grep { ! exists $slice_idx{$_} } 0 .. $#array; # Inverse slice. my @invslice = @array[@invslice_idx]; use Data::Dump 'dd'; dd { 'idx' => \@invslice_idx, 'invslice' => \@inv +slice }; my @newinvslice = @array; @newinvslice[@slice_idx] = (); @newinvslice = grep defined, @newinvslice; use Data::Dump 'dd'; dd { 'newinvslice' => \@newinvslice }; use List::Util qw( none ); my @new2invslice = @array[ grep { my $idx = $_; none { $idx == $_ } @slice_idx } 0 .. $#array ] +; use Data::Dump 'dd'; dd { 'new2invslice' => \@new2invslice }; my @new3invslice = @array; splice @new3invslice, $_, 1 for reverse @slice_idx; use Data::Dump 'dd'; dd { 'new3invslice' => \@new3invslice }; use List::Util qw( uniq ); my @new4invslice = @array[ (uniq @slice_idx, 0 .. $#array)[@slice_idx .. $#array ] ]; use Data::Dump 'dd'; dd { 'new4invslice' => \@new4invslice }; [download] Outputs: { array => ["aa" .. "zz"], slice => ["ag", "an", "bq", "co", "cr"], slice_idx => [6, 13, 42, 66, 69], } { idx => [0 .. 5, 7 .. 12, 14 .. 41, 43 .. 65, 67, 68, 70 .. 675], invslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], } { newinvslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], } { new2invslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], } { new3invslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], } { new4invslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], } [download]	[reply] [d/l] [select]
Re^2: Inverse slices by tybalt89 (Monsignor) on Nov 01, 2024 at 22:03 UTC
Found another one... `use List::AllUtils qw( singleton ); my @new6invslice = @array[ singleton 0 .. $#array, @slice_idx ];` [download] If that's not better, we need a better definition of 'better'. (WARNING: possible infinite recursion encountered!)	[reply] [d/l]
Re^2: Inverse slices by LanX (Saint) on Nov 01, 2024 at 15:29 UTC
Some suggestions, while I wait for Ubuntu to upgrade... My comments in all CAPS. Most importantly - regarding `@newinvslice` - Ikegami was right to point out that an `undef` value is not a sufficiently safe criteria (AKA the semi-predicate problem) I have a solution using the ref of a scalar var which MUST be unique. A bit clumsy, because Perl has no === operator, but I am open for more elegant suggestions. :) #!/usr/bin/perl use strict; # https://perlmonks.org/?node_id=11162546 use warnings; use Data::Dump 'dd'; use Test::More; $SIG{__WARN__} = sub { die @_ }; # Is there a better way to get an inverse slice than to do something l +ike: my @array = ('aa'..'zz'); my @slice_idx = (6,13,42,66,69); # Slice my @slice = @array[@slice_idx]; my $expected = [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ]; dd { 'array' => \@array, 'slice_idx' => \@slice_idx, 'slice' => \@slice, 'expected' => $expected, }; { my $TEST = "not in idx hash"; my %slice_idx; @slice_idx{@slice_idx} = (); # NO MAP NEEDED my @invslice_idx = grep { ! exists $slice_idx{$_} } 0 .. $#array; # Inverse slice. my @invslice = @array[@invslice_idx]; is_deeply(\@invslice, $expected, $TEST ) or diag dd { 'idx' => \@invslice_idx, 'invslice' => \@invslice }; } { my $TEST = "not delete value"; my @newinvslice = @array; my $delete; @newinvslice[@slice_idx] = (\$delete) x @slice_idx; # UNIQUE REF V +ALUE @newinvslice = grep { ref $_ ne "SCALAR" or $_ ne \$delete } @ne +winvslice; is_deeply(\@newinvslice ,$expected, $TEST) or diag dd { 'newinvslice' => \@newinvslice }; } { my $TEST = "splice backwards"; my @new3invslice = @array; splice @new3invslice, $_, 1 for sort { $a < $b } @slice_idx; # ENS +URE IDX SORTED is_deeply(\@new3invslice, $expected, $TEST) or diag dd { 'new3invslice' => \@new3invslice }; } done_testing; [download] `{ array => ["aa" .. "zz"], expected => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], slice => ["ag", "an", "bq", "co", "cr"], slice_idx => [6, 13, 42, 66, 69], } ok 1 - not in idx hash ok 2 - not delete value ok 3 - splice backwards 1..3` [download] Cheers Rolf _{(addicted to the Perl Programming Language :) see Wikisyntax for the Monastery} Updates Bugfix s/and/or/	[reply] [d/l] [select]
Re^3: Inverse slices by tybalt89 (Monsignor) on Nov 01, 2024 at 17:21 UTC
How to fix the semi-predicate problem using undef :) #!/usr/bin/perl use strict; # https://perlmonks.org/?node_id=11162546 use warnings; $SIG{__WARN__} = sub { die @_ }; # Is there a better way to get an inverse slice than to do something l +ike: my @array = ('aa'..'zz'); my @slice_idx = (6,13,42,66,69); # Slice my @slice = @array[@slice_idx]; use Data::Dump 'dd'; dd { 'array' => \@array, 'slice_idx' => \@slice_idx, 'slice' => \@slice }; my %slice_idx = map { $_ => 1 } @slice_idx; my @invslice_idx = grep { ! exists $slice_idx{$_} } 0 .. $#array; # Inverse slice. my @invslice = @array[@invslice_idx]; use Data::Dump 'dd'; dd { 'idx' => \@invslice_idx, 'invslice' => \@inv +slice }; my @wanted = 0 .. $#array; @wanted[@slice_idx] = (); my @new6invslice = @array[ grep defined, @wanted ]; use Data::Dump 'dd'; dd { 'new6invslice' => \@new6invslice }; [download] Outputs: `{ array => ["aa" .. "zz"], slice => ["ag", "an", "bq", "co", "cr"], slice_idx => [6, 13, 42, 66, 69], } { idx => [0 .. 5, 7 .. 12, 14 .. 41, 43 .. 65, 67, 68, 70 .. 675], invslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], } { new6invslice => [ "aa" .. "af", "ah" .. "am", "ao" .. "bp", "br" .. "cn", "cp", "cq", "cs" .. "zz", ], }` [download]	[reply] [d/l] [select]
Re: Inverse slices by ysth (Canon) on Nov 01, 2024 at 01:04 UTC
That works. You could make it more terse, if you wanted: `my @array = ('aa'..'zz'); my %slice_idx = map( ($_ => 1), 6,13,42,66,69 ); my @invslice = @array[ grep ! exists $slice_idx{$_}, 0..$#array ];` [download] Or use Array::Set, or Set::Array, or Set::Scalar. -- A math joke: r = \| \|csc(θ)\|+\|sec(θ)\| \|-\| \|csc(θ)\|-\|sec(θ)\| \|	[reply] [d/l]
Re^2: Inverse slices by choroba (Cardinal) on Nov 01, 2024 at 01:29 UTC
Or even `my %slice_idx; @slice_idx{6, 13, 42, 66, 69} = ();` [download] `map{substr$_->[0],$_->[1]\|\|0,1}[\\|\|{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^ARGV,3]`	[reply] [d/l] [select]
Re: Inverse slices by ikegami (Patriarch) on Nov 01, 2024 at 00:59 UTC
Better how? Shorter? Put the code in a sub. More readable? Put the code in a sub. Faster? Sort the indexes and use a merge sort-like approach written in C to filter out the unwanted elements. Alternative approach if the array only has defined elements: `my @invslice = @array; @invslice[ @slice_idx ] = (); @invslice = grep defined, @invslice;` [download] Update: Replaced `undef @invslice[ @slice_idx ];` (which doesn't work) with `@invslice[ @slice_idx ] = ();`.	[reply] [d/l] [select]
Re^2: Inverse slices by ysth (Canon) on Nov 01, 2024 at 01:06 UTC
While this use of undef works in all perl 5 versions I am aware of, it is in fact not documented. I prefer `@invslice[ @slice_idx ] = ();` [download] -- A math joke: r = \| \|csc(θ)\|+\|sec(θ)\| \|-\| \|csc(θ)\|-\|sec(θ)\| \|	[reply] [d/l]
Re^3: Inverse slices by Anonymous Monk on Nov 19, 2024 at 00:23 UTC
While this use of undef works in all perl 5 versions I am aware of, it is in fact not documented Really? It didn't work on any perl version I tried. For me, ikegami's slice undef trick only undefined the last element in the slice.	[reply]
Re^4: Inverse slices by ikegami (Patriarch) on Nov 19, 2024 at 17:51 UTC
Re: Inverse slices by johngg (Canon) on Nov 01, 2024 at 11:59 UTC
This uses a hash for the index rather than an array and push with a ternary. `johngg@aleatico:~$ perl -Mstrict -Mwarnings -MData::Dumper -E 'say q{} +; my @arr = ( q{a} .. q{m} ); my %idx = map { $_ => 1 } 2, 7 .. 9, 12; my @slc; my @inv; push @{ $idx{ $_ } ? \ @slc : \ @inv }, $arr[ $_ ] for 0 .. $#arr; print Data::Dumper->Dumpxs( [ \ @arr, \ @slc, \ @inv ], [ qw{ arr slc inv } ] );' @arr = ( 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm' ); @slc = ( 'c', 'h', 'i', 'j', 'm' ); @inv = ( 'a', 'b', 'd', 'e', 'f', 'g', 'k', 'l' );` [download] I hope this is of interest. Update:* Removed erroneous space in `Dumpxs` call which was mucking up the output. Cheers, JohnGG	[reply] [d/l] [select]
Re: Inverse slices by oodler (Beadle) on Nov 01, 2024 at 17:21 UTC
splice with map could work, but the better approach if possible would be to catch it on the front end, and filter it; you may make optimizations once you get it working for your purposes. A possible improvement would be to replace the grep with List::Util's `any` function. `use strict; use warnings; use Data::Dumper; my @skip_idx = (6,13,42,66,69); my @array = (); my @invarray = (); my $idx = -1; foreach my $str ('aa'..'zz') { ++$idx; if (grep { /^$idx$/ } @skip_idx) { # could use List::Util::any here push @invarray, $str; print "skipped '$str' (index of $idx)\n"; } push @array, $str; } print Data::Dumper::Dumper(\@invarray);` [download] Output, `$ perl test.pl skipped 'ag' (index of 6) skipped 'an' (index of 13) skipped 'bq' (index of 42) skipped 'co' (index of 66) skipped 'cr' (index of 69) $VAR1 = [ 'ag', 'an', 'bq', 'co', 'cr' ];` [download]	[reply] [d/l] [select]

Updates