Class MapReduce
Implements a simplistic version of the popular Map-Reduce algorithm. Acts like an iterator for the original passed data after each result has been processed, thus offering a transparent wrapper for results coming from any source.
Property Summary
-
$_counter protected
intCount of elements emitted during the Reduce phase
-
$_data protected
iterableHolds the original data that needs to be processed
-
$_executed protected
boolWhether the Map-Reduce routine has been executed already on the data
-
$_intermediate protected
arrayHolds the shuffled results emitted from the map phase
-
$_mapper protected
callableA callable that will be executed for each record in the original data
-
$_reducer protected
callable|nullA callable that will be executed for each intermediate record emitted during the Map phase
-
$_result protected
arrayHolds the results as emitted during the reduce phase
Method Summary
-
__construct() public
Constructor
-
_execute() protected
Runs the actual Map-Reduce algorithm. This iterates the original data and calls the mapper function for each record, then for each intermediate bucket created during the Map phase call the reduce function.
-
emit() public
Appends a new record to the final list of results and optionally assign a key for this record.
-
emitIntermediate() public
Appends a new record to the bucket labeled with $key, usually as a result of mapping a single record from the original data.
-
getIterator() public
Returns an iterator with the end result of running the Map and Reduce phases on the original data
Method Detail
__construct() ¶ public
__construct(iterable $data, callable $mapper, callable|null $reducer = null)Constructor
Example:
Separate all unique odd and even numbers in an array
$data = new \ArrayObject([1, 2, 3, 4, 5, 3]);
$mapper = function ($value, $key, $mr) {
$type = ($value % 2 === 0) ? 'even' : 'odd';
$mr->emitIntermediate($value, $type);
};
$reducer = function ($numbers, $type, $mr) {
$mr->emit(array_unique($numbers), $type);
};
$results = new MapReduce($data, $mapper, $reducer);Previous example will generate the following result:
['odd' => [1, 3, 5], 'even' => [2, 4]]Parameters
-
iterable$data The original data to be processed.
-
callable$mapper the mapper callback. This function will receive 3 arguments. The first one is the current value, second the current results key and third is this class instance so you can call the result emitters.
-
callable|null$reducer optional the reducer callback. This function will receive 3 arguments. The first one is the list of values inside a bucket, second one is the name of the bucket that was created during the mapping phase and third one is an instance of this class.
_execute() ¶ protected
_execute(): voidRuns the actual Map-Reduce algorithm. This iterates the original data and calls the mapper function for each record, then for each intermediate bucket created during the Map phase call the reduce function.
Returns
voidThrows
LogicExceptionif emitIntermediate was called but no reducer function was provided
emit() ¶ public
emit(mixed $val, mixed $key = null): voidAppends a new record to the final list of results and optionally assign a key for this record.
Parameters
-
mixed$val The value to be appended to the final list of results
-
mixed$key optional An optional key to assign to the value
Returns
voidemitIntermediate() ¶ public
emitIntermediate(mixed $val, mixed $bucket, mixed $key = null): voidAppends a new record to the bucket labeled with $key, usually as a result of mapping a single record from the original data.
Parameters
-
mixed$val The record itself to store in the bucket
-
mixed$bucket the name of the bucket where to put the record
-
mixed$key optional An optional key to assign to the value
Returns
voidgetIterator() ¶ public
getIterator(): TraversableReturns an iterator with the end result of running the Map and Reduce phases on the original data
Returns
TraversableProperty Detail
$_executed ¶ protected
Whether the Map-Reduce routine has been executed already on the data
Type
bool$_mapper ¶ protected
A callable that will be executed for each record in the original data
Type
callable$_reducer ¶ protected
A callable that will be executed for each intermediate record emitted during the Map phase
Type
callable|null