DBIx::Class::ResultSet::SetControl - Easier Looping over resultsets
Given a DBIx::Class::ResultSet that consumes this component, such as the following:
package MySchema::ResultSet::Bar; use strict; use warnings; use parent 'DBIx::Class::ResultSet'; __PACKAGE__->load_components('ResultSet::SetControl'); ## Additional custom resultset methods, if any 1;
Then later when you have a resultset of that class:
my $rs = $schema->resultset('Bar');
You can call methods directly on your object which are related to control flow and looping over the items in you resultset.
$rs->tap(sub { print shift->find({id=>1} ? 'found one' : 'nope'; })->each(sub { my ($each, $row) = @_; print $each->is_odd ? $row->name . ' is odd' : 'nope, not odd'; });
NOTE If you intend to use this component in many of your resultsets, its best practice to write a 'base' resultset that loads this (and any other) components from which all your custom resultset classes inherit.
There are times where Perl's procedural syntax for control flow and looping leads to excessively verbose code. For those times we present this helper which is designed to encapsulate some very common control flow and loop patterns for DBIx::Class users.
The methods are OO in nature and designed to be compact and concise.
Additionally, we have tried to write these methods to allow for a 'chaining' approach that you can't replicate with traditional Perl control and looping structures. Each control flow method returns the original resultset so you can proceed as though it is unaltered (unless of course you alter it somehow like with an insert or update).
The goal it to help avoid excessive conditional logic and to allow one to write more compact and neat code. For example, you could replace:
my $has_rows; while(my $row = $rs->next) { $has_rows = 1; ## Do something } unless($has_rows) { warn 'no rows!'; }
With something like
$rs->each(sub { my ($each, $row) = @_; ## Do Something }, sub { warn 'no rows!'; });
The second version has less overall lines and characters, and it also carefully encapsulates a very common pattern, which is to loop over all the rows in a resultset and do something should no rows exist. Also, the "each" method returns the original $rs so you could chain commands:
$rs
$rs->each(sub { my ($each, $row) = @_; ## Do Something }, sub { warn 'no rows!'; })->tap(sub { my $rs = shift; ## Do something else });
There may be cases in your logical flow where this type of programming is more clear and simple; in other cases traditional Perl control and looping might be better. These methods give you an option. On the other hand you might think this is all pointless line noise. As you wish :)
This component defines the following methods.
Arguments: $rs->map(\my @mapped_to, $coderef, ?$if_empty_coderef) Returns: Original Resultset OR partly iterated Resultset
Example:
$schema ->resultset('Person') ->map(\my @list, sub { my ($i, $row) = @_; return +{ id => $row->id }; });
Where @list will be an array of hashes in the form +{ id => 111 }.
Basically this is like:
my @list = map { ${ id => $_->id } } $person_rs->all;
The two main advantages are that the method returns the resultset so you can chain it and the $i iterator has some ability to control and escape the loop. Also ->all will inflate the entire set whereas map will inflate one at a time (possible memory savings.)
$i has some nice 'status of the loop' features such as if the current $row is the first one, or if its odd/even. Be sure to check DBIx::Class::Util::ResultSet::Iterator.
It also has an optional third argument that only runs if the set is empty. It runs once and can be used to provide a default value to the target array:
$schema ->resultset('Person') ->map(\my @list, sub { my ($i, $row) = @_; return +{ id => $row->id }; }, sub { return 'empty list'; });
Its not amazingly useful standalone but if you are using the other stuff its a nice to have to make the API more complete.
Arguments: $rs->each($coderef|\@$coderef, ?$if_empty_coderef) Returns: Original Resultset OR partly iterated Resultset
Where $coderef is an anonymous subroutine or closure that will get the instantiated DBIx::Class::Helpers::Util::ResultSet::Iterator object and the current $row from the set returned. For example $row in the ResultSet the $coderef will be executed once.
$coderef
$row
$if_empty_coderef is an anonymous subroutine or closure that gets executed ONLY if there were no rows in the set. It gets the $resultset as an argument (this might change later if we discover a better thing to do here).
$if_empty_coderef
$resultset
In the case where the first argument is an arrayref of coderefs, we automatically iterate over each coderef for each result in the set in turn and reset the coderef iterator as needed to make sure we hit every item in the set. Please be aware that in the case where the arrayref of coderefs is longer than the available results in the set, this means that not all coderefs will be invoked and this happens without an exception being thrown.
Example: For the given DBIx::Class::ResultSet, iterator over each result.
$rs->each(sub { my ($each, $row) = @_; ... });
This is functionally similar to something like:
my $itr = DBIx::Class::Util::ResultSet::Iterator->new(resultset=>$rs); while(my $row = $itr->next) { ... }
However the method will return the original $resultset used to initialize it so that you can continue chaining or building off it. Of course you will need to issue a c<ResultSet->reset> for this to be useful.
Here's a more detailed example.
$rs->each(sub { my ($each, $row) = @_; $each->first(sub { print "Hey, this is the first row!"; }); if($each->is_odd) { print $row->columnname; } else { return $each->escape; } }, sub { my ($rs) = @_; warn "The resultset was empty, nothing done..."; });
Finally one example using an arrayref as the first argument:
$rs->each( [ sub { ... }, sub { ... }, sub { ... }, ], sub { my ($rs) = @_; warn "The resultset was empty, nothing done..."; } );
You may find this helper leads you to writing more concise and compact code. Additionally having an iterator object available can be helpful, particularly when you are in a template and need to display things differently based on if the row is even/odd, first/last, etc.
You should see the documentation for DBIx::Class::Util::ResultSet::Iterator for the methods this object exposes for use.
NOTE For conciseness in simple cases, we overload $_ to equal the value of $row as described above.
$_
Arguments: $rs->once($coderef, ?$if_empty_coderef) Returns: Partly iterated Resultset
Works just like "each" expect instead of iterating over the entire resultset we just take the first $row.
Example
$rs->once( sub { my ($row) = @_; }, sub { warn 'no rows left!'; })->each( ... );
Useful to isolate the logic for the first row in a resultset.
Arguments: $coderef, ?@args Returns: Original Resultset
Do a coderef with the resultset passed as an argument.
$rs->tap(sub { my ($func, $rs, $arg) = @_; $rs->find({id=>$arg}); }, 100);
If you pass more than one argument, all the extra arguments will be send to the anonymous coderef.
NOTE The resultset passed is cloned from the original and it returns the original resultset so you can chain.
NOTE the first argument passed to the coderef is the orginal coderef so that you can call recursively.
NOTE For conciseness in simple cases, we overload $_ to equal the value of $rs as described above.
Arguments: $integer, $coderef, ?@args Returns: Original Resultset
Basically this calls "tap" a number of times equal to the first argument.
$rs->times(3, sub { my $rs = shift; ... });
John Napiorkowski email:jjnapiork@cpan.org
DBIx::Class
Copyright 2017, John Napiorkowski email:jjnapiork@cpan.org
This library is free software; you can redistribute it and/or modify it under the same terms as Perl itself.
To install DBIx::Class::ResultSet::SetControl, copy and paste the appropriate command in to your terminal.
cpanm
cpanm DBIx::Class::ResultSet::SetControl
CPAN shell
perl -MCPAN -e shell install DBIx::Class::ResultSet::SetControl
For more information on module installation, please visit the detailed CPAN module installation guide.